Spotted Antbirds

THE BEHAVIOR OF SPOTTED ANTBIRDS

EDWIN O. WILLIS

ORNITHOLOGICAL MONOGRAPHS NO. 10

PUBLISHED BY

THE AMERICAN ORNITHOLOGISTS' UNION

1972 Above: Male Spotted Antbird BX, who occupied a territory near the center of Barro Colorado Island for at least the years 1960-1971. Below: Mist-netting a male Spotted Antbird. Bands, read up the left leg and down the right, identify this individual as male CWRS (orange-red, white, red, black-yellow). THE BEHAVIOR OF SPOTTED ANTBIRDS

EDWIN O. WILLIS

ORNITHOLOGICAL MONOGRAPHS NO. 10

PUBLISHED BY

THE AMERICAN ORNITHOLOGISTS' UNION

1972 ORNITHOLOGICAL MONOGRAPHS

This series,published by the American Ornithologists'Union, has been establishedfor major paperstoo long for inclusionin the Union's journal, The Auk. Publicationhas been made possiblethrough the generosityof Mrs. Carll Tucker and the Marcia Brady Tucker Foundation,Inc.

Correspondenceconcerning manuscripts for publicationin the series shouldbe addressedto the Editor, Dr. Robert M. Mengel, Museum of Natural History, University of Kansas,Lawrence, Kansas 66044.

Copiesof OrnithologicalMonographs may be orderedfrom the Treasurer of the AOU, Butt L. Monroe,Jr., Box 23447, Anchorage,Kentucky 40223. (See price list on insideback cover.)

OrnithologicalMonographs, No. 10, vi + 162 pp. Editor, Robert M. Mengel Editorial Assistant,James W. Parker

Issued August 29, 1972 Price $6.00 prepaid ($4.75 to AOU Members)

Library of CongressCatalogue Card Number 72-87186 Printedby the Allen PressInc., Lawrence,Kansas 66044 TABLE OF CONTENTS

INTRODUCTION ...... 1 Spotted Antbirds and Their Relatives ...... 1 Army Ants ...... 3 Study Areas ...... 3 Methods of Study ...... 4

ACKNOWLEDGMENTS ......

GENERAL BEHAVIOR ...... 6 Calls and Songs ...... 6 Normal Postures and Movements ...... 9 Resting ...... 13 Preening and Head-Scratching...... 13 Stretching ...... 14 Egestionand Drinking ...... 14 Sunning ...... 15 Bathing ...... 15 Anting ...... 15 Attacks by Arthropods ...... 16 Reactions to Rain ...... 16 Discussion ...... 16

REACTIONS TO DANGER ...... 18 Freezing ...... 18 Chipping and Panicking...... 19 Chitring and Mobbing ...... 20 Fleeing and Tameness...... 23 Investigating ...... 24 Reactionsof Captives...... 24 Discussion ...... 24

AGONISTIC BEHAVIOR ...... 29 Submissiveand Escape Behavior ...... 29 AggressiveDisplays and Attack Behavior ...... 30 Sequencesof Agonistic Behavior ...... 34 Discussion ...... 36

REPRODUCTIVE BEHAVIOR ...... 39 Singing and Wandering ...... 39 Chirping and Flirting ...... 39 CourtshipFeeding ...... 40 Grooming ...... 42 Pair Formation ...... 44 Copulation ...... 46 The Nest Site and Nest ...... 48 Building the Nest ...... 49 The Eggs and Incubation ...... 51 Growth of the Young ...... 57 Fledglings ...... 59 Distraction Displays and Parental Alarm ...... 63 Development of Young out of the Nest ...... 65 Renesting and Nest Success...... 67 Breeding Season...... 70 Discussion ...... 7 !

WANDERING YOUNG AND TERRITORIAL ADULTS: SPATIAL BEHAVIOR ...... 83 Dominance and Territoriality ...... 87 Territory Size, Densities, and Biomasses...... 89 Habitat Selection ...... 89 Discussion ...... 90

FORAGING BEHAVIOR ...... 96 Perch Selection ...... 97 Foraging Away from Ants ...... 99 Searchingfor Army Ants ...... 100 Following the Army Ants ...... 101 Prey and Prey Treatment ...... 103 Numbers at Swarms ...... 106 Competition at Swarms ...... 109 Discussion ...... 116

ASSOCIATION WITH INTERSPECIFIC FLOCKS ...... 123 Other Members of Flocks ...... 126 CompetitionAmong Flocking Birds ...... 133 Discussion ...... 135

SUMMARY ...... 148

GAZETTEER OF LOCALITIES ...... 151

LITERATURE CITED ...... 152

APPENDIX 1 ...... 156

APPENDIX 2 ...... 158

vi INTRODUCTION SpottedAntbirds (Hylophylaxnaevioides, Formicariidae) are commonbirds of the undergrowthof lowlandtrc•pical forests from Hondurasto Ecuador. Skutch (1946, 1969), Johnson (1954), and Slud (1960) studied them briefly, but did not investigatethem in detail. Since 1960, as part of a study of ant-followingbirds on Barro ColoradoIsland in the PanamfiCanal Zone (Willis, 1967), I havestudied the behaviorof SpottedAntbirds. This report detailsthe resultsof that study. SpottedAntbirds (frontispiece) are ir}teresting because they are regular members of the two major typesof mixedbird flocksthat have attractedthe attentionof naturalistsin tropicalforests since the time of Bates(1863). These two typesof flocksare thosethat follow army ants (Chapin,1932; Johnson, 1954) andwandering flocks (Swynnerton, 1915; Winterbottom,1943, 1949; Davis, 1946; Stanford,1947; Rand, 1954; Short, 1961; Moynihan,1962a; McClure,1967). The flocksaround army antsgather primarily to capture insectsflushed by the ants,and thus are no•t very different from aggregations of birdsat fruitingtrees, garbage dumps, and otherconcentrated sources of food. Wanderingmixed flocks seldom concentrate at local food sources,so that the advantagesof flockingare currentlyunknown. Morse (1967) and othersfavor various theories that the birdsgain food advantagesby flocking, but Moynihan(1962a) andothers think that flocking somehow reduces predation. Most authors have had little time for careful studies of individual species,and have basedtheir theorieson tabulationsof all the individualsand speciesin observedflocks. The studyof SpottedAntbirds gives information froma newdirection, and suggests new approaches for the studyof wandering flocks. After the presentintroduction, there are sevensections in thisreport: these treatgeneral behavior, antipredator behavior, agonistic behavior, reproductive behavior,spatial behavior, foraging behavior, and flockingbehavior. There is a discussion at the end of each of these seven sections.

SPOTTED ANTBIRDS AND THEIR RELATIVES Male SpottedAntbirds (frontispiece)are brightlypatterned with white, chestnut,dark gray,and black. The namecomes from a necklaceof prominent blackspots across the chest, between the conspicuously white upper breast and the whiteto grayish-whitebelly. The headis dark gray,contrasting slightly with the blackthroat and bill and with the dark rufouseyes. The back is chestnut,as are two broad and conspicuous bars on theblack wings. The lesser covertsare tipped with white speckles, while the dark remigesare tippedor crossedby a third wingbar of buff to chestnuthue. White basesto the central backfeathers form a largedorsal patch that is usuallyconcealed, except when 2 ORNITHOLOGICAL MONOGRAPHS NO. 10 birds dispute. Each brown tail feather has a black subterminalband and a buffy-whitishtip. The femmeSpotted Antbird has bright buff to chestnutbands and feather tippingson the blackishwings, but is otherwisemore soberlycolored than is the male. She is brownish-chestnutabove, with a concealedwhitish dorsal patchthat can be displayedin disputes.The tail hasa buffytip and a brownish- black subterminalband. The head is brown, fading to buffy-whitishon the indistinctsuperciliary line and lower face and to white on the throat. The whitishunderparts are tinged with pale buff, especiallyon the sides,and thereis a more or lessobscure necklace or high breastband of brownishspots. The bill and legsare dark gray and the eyesdark whitishgray. ! have seenonly one partial albino, a male with much white on the wrist area on Barro ColoradoIsland on 22 July 1967. Spotted Antbirds range in Caribbean lowland forests from southeastern Hondurasto the lower MagdalenaValley of northernColombia. Over the Cordillerade Guanacasteof northernCosta Rica (Slud, 1964) and in central Panama (El Valle) they rangeonto the Pacific slope. From centralPanama to central Ecuador they occur in most of the forestedPacific lowlands. I have found them as high as 900 m elevationon Cerro Campanain central PanamA,and there are specimensin the United StatesNational Museum from as high as 1,300 m on Cerro Tacarcuna,eastern PanamA. East of the Andes live two very similar species,the Spot-backedAntbird and the Dot-backed Antbird. (Common names used in this paper are from Meyer de Schauensee,1966'; the correspondingscientific names, except for speciesmentioned in papers cited, are given in the index.) The latter is a flycatching,timid little bird of the vdrgea,or periodicallyflooded woodlands alongrivers. It behavesand calls very differentlyfrom the SpottedAntbird, despiteits similar appearance. The Spot-backedAntbird behavesmore like the Spotted Antbird, but flycatchesin and near foliage more often and followsarmy antsless often. (Skutch [1946: 18] reportedthat an Ecuadorian speciesof Hylophylax, probably the Spot-backedAntbird on geographical and behavioral grounds, is much like the Spotted Antbird.) Since Spot- backed and Dot-backed Antbirds are similar morphologicallybut occur togetherwithout interbreeding, the similar SpottedAntbirds might also fail to interbreedwith Spot-backedAntbirds were they to comein contact. Among related antbirds,sympatric species of a genusor subgenuscommonly differ as litfie as thesetwo in behavior or morphologyor both. For thesereasons,

* While many of these names may not now be truly "common," in the senseof having wide currency, it seemsto me that this is to be hoped for and will be promoted by using them. PLATE1. Army ants (Eciton burchelli). Upper left, a major ("soldier") and a worker Upper right. a dense swarm of army ants starting to raid. Lower left, army ants carrying prey from the swarm to the bivouac. Lower right, corner of a nomadic bivouac made of bodies of ants and enclosing the queen and larvae. 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 3

I shallfollow traditionaltaxonomic practice and regardSpotted and Spot- backedantbirds as separatespecies, even though they are closelyrelated.

ARMY ANTS Rettenmeyer (1963), Schneirla (1957), and Willis (1967) summarize information on the army ants followed by birds. The two important ant speciesare Eciton burchelli and Labidus praedator. These ants form wide raidingswarms, flowing by the thousandsover the leaf litter and into tangles near or abovethe ground,in tropical to subtropicalforests from M•xico to Argentina. There are many speciesof birdsthat follow the antsand snapup arthropodsfleeing from the advancingarmies. No bird, so far as is known, regularlyeats the antsthemselves. Eciton burchelli (Plate 1), brown-and-yellowants that averagealmost a centimeterin length, swarm predictablyand in the daytime all year long. These large ants flush many large arthropods,such as spiders,roaches, crickets, and katydids. Labidus praedator, black ants averagingabout five mm long, flush more small prey, such as sowbugsand amphipods. They swarmabove ground mainly in rainy weather,and swarmat any hour of day or night. A raid of L. praedatoris likely to disappearunderground after a few hours and leave ant-followingbirds stranded. Those ant-following birds that follow army antsmore than 50 percentof the time usuallyfollow E. burchellirather than the unpredictableL. praedator. Birds that follow army antsless persistently--including many migrantsas well as the Spotted Anthird•follow both ant species(Willis, 1966a: 211). There are a few other army antsthat birds sometimesfollow. Nomamyrmex esenbecki,brown-and-yellow ants that resembleEciton burchelli,occasionally form stragglingswarms on the leaf litter in Neotropical forests. Five times on Barro Colorado I recordedSpotted Antbirds attendingraids of N. esenbecki. Once on Barro ColoradoI found a pair of SpottedAntbirds at a looseevening swarm of Eciton mexicanurn,a mainly nocturnalspecies that lookslike a smallE. burchelli.Once I foundSpotted Antbirds at a straggling swarmof an unknownspecies of Eciton On Barro Colorado,and onceat a swarmof anotherunknown species on BuenavistaPoint near Barro Colorado. Althoughcolumn-raiding army ants, especiallyEciton hamatum,are very commonon Barro Coloradoand in otherareas where I have studiedSpotted Antbirds, I have never seen SpottedAntbirds show interestin them. The columnraiders, the raiderswith stragglingswarms, and the semi-nocturnal raidersamong army ants are unlikely to be importantfor diurnalant-following birds. STUDY AREAS I studiedSpotted Antbirds at severallocalities in Panatari and Colombia (Appendix 1). All localitiesbut Yuto and Tanand6 (Choc6, Colombia) 4 ORNITHOLOGICAL MONOGRAPHS NO. 10 are characterizedby tropicalor lower subtropicalforests, moderately hot and humid,and havemarked dry seasonsbetween December and April. In the Choc6,where there are heavyrains all year,Spotted Antbirds are lesscommon than in the other areas. Most of my studieshave been on Barro ColoradoIsland, the research stationof the SmithsonianTropical Research Institute in the PanamaCanal Zone. This islandis a 15.7 squarekm hilltopwhich was isolated from nearby lowland forests when the waters of Gatun Lake rose, between 1911 and 1914, to form the centralpart of the nearbyPanama Canal. It is coveredby forest except at the laboratoryclearing. There is a good systemof trails, markedat 100-meterintervals by posts,so that one can easilyand accurately map the locationsof birds seen. In addition,I set up a grid of compass4ine trails near the center of the island. I visitedthe island as follows: 28 September1960-25 November 1961; 18 January-18February and 25 June-1 August 1962; 20 June-31 August 1963; 18 June-1 September1964; 15 January-5 March and 4 August-ll October 1965; 16 May-31 July 1966; 1 June-16 August 1967; 14 June- l September1968; 28 June-27 August 1969; 27 June-7 September1970; and 17 December1970-25 January1971. The geologyof Barro Coloradois discussedby Woodring (1958), and the climateand vegetationby Kaufmann (1962), Willis (1967), and earlier authors. Rainfall averages2,730 mma year, rising to a monthly high of 454 mm in Novemberand then droppingsharply, so that only 7.8 percentof the yeafly total falls from Januaryto April. Much of the easternhalf of the island and someof the westernhalf were in clearingsor low secondgrowth in 1923, when the island was set aside as a biological reserve. However, the fairy mature forestsin these areas are now rather like the older forests of the island, althoughsomewhat lower and with fewer treefalls and other clutter in the undergrowth.Little light reachesthe lower levels of the undergrowth (Allee, 1926: 288-289), and the annualdry seasoninhibits epiphytes and lush undergrowth,so that the lower layers of the forest are mostly open and uncluttered. Thin saplingsand spindly sproutsand palmetto reach upward for light, while buttressedtrees and clumps of palms (Oenocarpus panamensis)block long vistas,but the forestis not the impenetrable"jungle" one finds in areas recently disturbed by man. Treefalls, dense sapling tanglesaround rotting old treefalls,and patchesof spiny-leaved"wild pine- apples"(Ananas magdalenae)add a mosaicpattern of densepatches to the lower levels of the forest in some areas.

METHODS OF STUDY SpottedAntbirds are beingstudied in muchthe sameway as were Bicolored Antbirds (Willis, 1967). Birds are capturedin mist nets set ahead of swarms 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 5 of ants or elsewherein the forest and marked for later recognitionwith different combinations of colored celluloid bands. In field notes and in this report, singleletters representa band. Bands are read up the left leg and down the right. Thus, the male Spotted Antbird RWBP had a red band below a white one on his left leg and a blue band above a pink one on his right leg. Between 1960 and 1971, I banded498 SpottedAntbirds on and near the study area on Barro Colorado. I watchSpotted Antbirds, at and awayfrom swarmsof ants,from distances of 5 to 20 m to minimizefear reactions.These birds readilybecome fairly tame, and forage near one without signsof alarm. I took still pictureswith an Asahi Pentax and a 200 mm Takumar lens, usingelectronic flash and Kodak High-speedEktachrome for color pictures and various films, from Plus-X to Tri-X, for black-and-whites. Movies were taken at 24 to 32 framesper secondon Tri-X film, usingmainly an Arriflex cameraand Anglenieuxlens. I recordedbird callswith a Mohawk "Midgetape500" recorderat 9.5 cmper second or on a Uher"4000 ReportS" at 19 cm per second.To bring birds closerfor recordingsor experiments during1961-62, I usedthe Midgetapeand a Victor XT-401 amplifierand speaker(Victor Companyof Japan,Ltd.). Calls were analyzedon Kay ElectricCompany "Sonagraphs," using the wide-bandfilter passes.

ACKNOWLEDGMENTS I appreciatethe supportof a Woodrow Wilson travelingfellowship, pre- doctoral fellowshipsfrom the National Science Foundation, postdoctoral fellowshipsfrom the Frank M. Chapman Fund of the American Museum of Natural History, and a SigmaXi researchgrant. The staff and facilitiesof the SmithsonianTropical Research Institute, Panamfi Canal Zone, and of the CompanlaMinera Choc6-Pacifico,Colombia, were welcomeand welcoming. GeorgeCox, Gordon Orians, Dennis Paulson,Susan Smith, and otherspro- vided vigorouscriticism for the manuscript,which is not to say that they approveall of its presentpositions and conclusions.I appreciatethe help of Richard Stallcupin the 1964 observationsand of StephenKistler in the 1970 observations.My wife, Yoshika Oniki, helpedwith observationsand in many other ways. GENERAL BEHAVIOR • CALLSAND SONGS SpottedAntbirds are very like BicoloredAntbirds (Willis, 1967) in calls, so I shall use the sameterms for the calls of both. SpottedAntbirds ap- parentlylack keening,grunting, and growling calls. Male andfemale Spotted Antbirds have similar calls and songs. Chirring.--A sharprattle or buzz,chi'i'i'i'i, is the usualreaction of Spotted Antbirdsto disturbinglarge animals,such as humans(Figure 1, f). The nine oscillationsin the graphedchirr last 0.23 secondand rangefrom 1,500 to 6,500 Hertz in frequency(abbreviation Hz = cps). The mainenergy of the call risesfrom about 4,000 Hz to 4,500 at the third oscillation,then drops graduallyto 3,500. Concurrently,the oscillations speed up to 40 per secondat thethird and slow to 30 per secondby theend. The faint high-frequencyclicks terminatingeach oscillation may be clickingof the bill. The call remarkably resemblesthe chirring(Figure 1, g) of the Dot-wingedAntwren, a bird that doesnot follow army ants but which associateswith SpottedAntbirds in interspecificflocks. The chirringof the Dot-wingedAntwren is somewhat low and slow,ranging from 1,500 to 5,500 Hz and with about30 oscillations per secondthroughout; it lasts0.28 secondfor nine or so oscillations. Chipping.--A loud, sharppeep! or seriesof severalsuch notes (Figure 1, h, k, 1) is the usualreaction when a SpottedAntbird is very excited,as when a hawk flies up. Two chips (1), graphedat half speedin h to show high frequencies,are simpleinverted-V cries 0.05 secondlong and 0.35 secondapart. The mainharmonic rises with onebrief lag from about2,500 Hz to 6,500, thendrops evenly to 2,500. A presumedchip (k) of a fledglingin the hand lacks a descendingpart; it is about 0.02 secondlong and risesfrom about 2,700 to 3,800 Hz, with a harmonic at about 7 kHz. Singing.---Formy definitionof songsee discussionon p. 72. The characteristic"peety weety" song (Eisenmann, 1952) goes beeeeeeee,tipee, tipeeti,peeti, peeti or the like (Figure 1, a). After the long (0.5 second) initial note, short (0.05 second) ti notes alternatewith quaveringpeeee

Figure 1. Wide-band sonogramsof calls of SpottedAntbirds (one call of Dot-winged Antwren), from tapesat 9.5 and 19 cm/sec. a, Loud-songof a male. b, Faint-songof a female calling fledglings,printed heavily so insect noise showsin background. c, Snarl. d, Bugling,with insect noisesabove 6 kHz. e, Four chirps. f, Chirring of Spotted Antbird. g, Chirring of Dot-winged Antwren. h, Two chips, at half speed. i, Loud- peepingof fledgling (peeee,]eeee), with insectnoise above 4 kHz. j, Faint-peepingof fledgling (piet wiet wiet wiet), with insectnoises in background.k, Chip of fledgling, in hand. 1, Same two chips as in h, played at normal speed. 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 7 8 ORNITHOLOGICAL MONOGRAPHS NO. 10 whistles(0.3 second) at about four notes per second. In a typical whistle thereare 10 or so oscillationsof a few hundredHz about an averagefrequency of some3 kHz. Each simpleti cry risesfrom about2,500 to 3,500 Hz, with a brief middle lag, and drops evenly to about 2,800 Hz. Up to 10 or 15 beeeti phrasesare in the loudest songs;"faint-songs" (Figure 1, b) are softerand seldomgo over five suchphrases. In comparisonwith the "loud- song" above, the faint-songillustrated shows shorter whistleswith stronger oscillationsand a more strongly rising frequency. "Serpentine-songs"are faint-songsrepeated again and again, punctuatedby series of chirping or peup notes, such as beeee tipee tipeeti, peup-peup-peup-peup, bee tipeeti, peup, peup-peup-peup.Birds serpentine-singmainly when about to feed matesor young;faint-songs and loud-songsgo betweenmates or from parents to young; loud-songsare used by opposingbirds, too. Faint-songsand serpentine-songsare very variable. Songsof females often seem weaker than songsof males,but I have not founddependable ways to tell the sex of a bird from its songalone. Chirping.--Bothmale and femaleutter faint peupnotes (Figure 1, e) when near each other. In copulation,the chirpsof the male deepento a pip, pip, pip series. One femalegave a faint meu or meuhhhhiss during copulation, perhapsa different note or perhapsfaint snarling. A chirp curves down from about 2,800 to 1,200 Hz over 0.08 second,and has a faint descending overtonefrom 3,500 to 2,200 Hz as an apostropheat the end. Snarling.--A long, hissingwrieeeeeeeeehhhh! (Figure 1, c) is hurled at an opponentwhen a bird performsthe agonisticdisplay of challenging.The first part (0.2 second) of the snarl drops quickly from 3,500 to 3,000 Hz and stays there while oscillatingat 75 per second;the partly overlapping secondpart (0.95 second) explodesin hissingcomplex noise, with multiple oscillationsat about the same rate, between 2,500 and 5,000 Hz; the third part (0.02 second)is a short whistlethat dropsfrom 2,500 to 2,000 Hz and has a harmonicabout 4 kHz. The completesnarl here is 1.12 secondslong. Bugling.--A sharp twit! or chwit! note (Figure 1, d) is often the first call whena bird jerks uprightto start a seriesof snarlingchallenges. It apparently is homologouswith bugling in Gymnopithys and Rhegmatorhina (Willis, 1967, 1968, 1969a), but is much shorter--about 0.15 second. The sinuous whistledrops from some 5,500 to 2,000 Hz via a brief lag about 4,500 Hz, risesto over 6,000 Hz and quicklydrops below 4,500 Hz. An overtoneis at 4,000 Hz for its low point; further details are not visible because the recordingis faint and its SOhogramblends with backgroundnoise. Hissingand Snapping.--A bird dartingpast anotherin a supplantingoften gives a hissingdzihht! or chihhht! or szapp! The note often ends with a sharpsnap of the bill, but separatesnapping is rare. Whimpering.-•A bird persistentlysupplanted by a dominant rival gives 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 9

Figure 2. SpottedAntbirds. a, Male, somewhatsleeked in freezing pose,on sloping sapling (from photograph). b, Female, somewhatin the panickingpose, perched near ground (from field sketch). c, Male lungingto peck at the ground (from field sketch). d, Female standson ground and looks about for prey that escaped(from field sketch). e, Male looks about (from photograph). faint peepingnotes, pt pee-pee-peeor the like. Adult, dominantmales and othersoften give this note in the handwhen banded or recaptured. Peeping.-•Youngbirds peep in severalways. "Loud-peeping"(Figure 1, i) is often a loud, two-notepeeee, jeeee! the last note at a lower pitch. The il- lustratedloud-peep descends throughout from about 3,600 to 3,400 Hz, is about 1.13 secondslong (first note 0.62, gap 0.13, secondnote 0.38), and oscillatesslightly at about 14 per second.Loud-peeping resembles the song of the Ruddy-tailedFlycatcher. "Faint-peeping" varies greatly, from a sibilant piet-wiet-wiet-wiet-wietthat becomeslike the songof the adult as the young bird grows,to a soft pee, pee-pee-pee(Figure 1, j) or a long, soft chieeh, chieehlike a faint versionof loud-peeping.At timesthere are faint wd grunts betweenfaint-peeps as a kind of "serpentine-peeping."The illustratedfaint- peepinghas slight irregular oscillations between 3,000 and 3,500 Hz; perhaps the oscillatorymechanism is not underfull controlat this age. Squeaking.•Young squeakscraihh series when fed. Screaming.--In the hand, birds occasionallyscream sharply. One female gave a shortwaiaiaihh! scream when supplantedby a Gray-headedTanager.

NORMAL POSTURES AND MOVEMENTS To describethe posturesor posesof an animal, it is convenientand in- structiveto depict a "normal" or "standard"posture, in which the animal is "just standing,"and to specifythe movementsof feathersand other parts of 10 ORNITHOLOGICAL MONOGRAPHS NO. 10

the body that producespecial poses (Orians and Christman,1968: 5-6; Willis, 1967: 16). The standardposture for an undisturbedSpotted Antbird (Figure 2, a, e) is much like that for a Bicolored Antbird (Willis, 1967: 16). Like the Bicolored Antbird, the Spotted Antbird is adept at clinging to slender vertical or inclined saplings (Figure 3). Both cling by flexing the upper leg, extendingthe lower leg, and anglingtoe 11 on the lower foot some 20 to 40 degreesabove the closelyappressed toes III and IV. The body is therebytilted or rolled toward the perch (Figure 2, b). "Flicking" the tail, or lowering it below the line of the body and jerking it suddenlyback to near that line, is one of the most characteristicmovements of SpottedAntbirds. Among antbirds,Hylophylax and related generaflick the tail, while antbirdsof Myrmeciza and related genera "pound" the tail emphaticallydownward and raise it slowly (Willis, 1967: 39). As is dis- cussedunder "reactionsto danger"and at other places,these tail movements usuallyindicate that the observeror other factorsare disturbingor exciting the birds,for tame onesthat are inactiveor preeningrarely flick or pound the tail. 1972 WILLIS: BEHAVIOR OF SPO'I-'I-FI) ANTBIRDS 11

, I

Figure 4. Spotted Antbirds jumping from nests. Male (above) is just opening wings; female (below) is jumping downward rather than flying.

The small SpottedAntbird (weighing 15 to 22 g) hops along limbs more easily and frequentlythan does the BicoloredAntbird (25 to 35 g). Both speciesreadily yaw and pitchon a horizontalperch or arounda verticalperch, pivot or reverseon or along a perch, and hop from perch to perch or on the 12 ORNITHOLOGICAL MONOGRAPHS NO. 10

Figure 5. Preening (a) and resting (b) female SpottedAntbirds, from field sketches. ground. They do so in suchstereotyped ways and after suchdistinct pauses that they sometimesremind one of clockworktoys. Despiteits smallsize, the SpottedAntbird generally jumps like the Bicolored Antbird to start flight. Figure 4 showsbirds jumping from a nest. The first part of the flight is often a parabolicarc, or an arc like that of a projectile, followedby a moreor lessstraight course to the next perch. The male (Figure 4, above) is startingto open his wings on a parabolicjump. Crippled or youngbirds flutter to start flight rather than jump strongly. Flight is fluttery, slow, direct, and generallymuch like that of BicoloredAntbirds. However, SpottedAntbirds hover and maneuverthrough dense tangles more easily than do the larger birds. If frightened, a Spotted Antbird flies as rapidly as a flushedquail. An alightingbird brakeswith its wingsas well as with its legs. Control, not speed,seems the keynoteof its flight. A traveling SpottedAntbird generallyflies 1-20 m at a time, at 1-2 m abovethe ground,inside the forest. One rarely seeslonger flights, higher traveling,and movementsoutside the forest or acrossclearings or large streams.I only oncesaw a bird crossthe laboratoryclearing on Barro Colo- rado, althoughseveral pairs have territoriesaround the clearingand come up to the bamboosand other vegetationwithin a meter or two of the edge. The one crossingwas in 1971, whenbamboos had grownup to providea shadedpassage. In generalthey avoidsunlit areas; even sunflecksrarely strikethem. On CerroCampana, I oncesaw them in openwoods from which all undergrowthhad beencleared; these birds were followingarmy ants near a house. A travellingbird alightseasily on eachvertical sapling along its course, looks aboutfrom the clingingposture, swings like a gate (pitches) around the perch, and flies to the next perch. Unlike BicoloredAntbirds, which seldomforage as they travel,wandering Spotted Antbirds often forageby 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 13

TABLE 1 RECORDS OF PREENING HEIGHTS FOR SPOTTED ANTBIRDS

Height in Meters 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 2 Total Number of Records 0 6 13 17 18 9 10 7 5 0 8 93 cockingor rotatingtheir headsto studythe groundor by dartingfor prey. A wandererseldom moves far in onedirection unless going to a nest;usually it circlesor winds about in an area. At times sharp chippingnotes or songs mark its course. RESTING A restingSpotted Antbird (Figure 5, b) generallysits on a horizontal perchnear the groundin or nearmoderately dense cover, especially in the tangledlimbs of fallen treesor in denselianas or under a palm clump (Oenocarpuspanamensis). The relativelyopen undergrowth where the ant- birdsforage is lesslikely to be usedfor resting,although short periods of restingsometimes interrupt foraging. Besides flexing the legs to sit, the restingantbird fluffs the ventralfeathers and back feathersuntil they cover the feet and sometimesthe wings,closes and lowers the tail, and raisesand retracts the neck so the head nestles on the shoulders. It looks around, but seldomcocks the head as if looking at the groundfor prey. Commonly preening interrupts resting.

PREENING AND HEAD-SCRATCHING Preening SpottedAntbirds generallytake horizontal perches, 1-3 cm in diameterand 0.2-0.9 m abovethe ground (Table 1) in the edgesof tree-

TABLE 2 PERCH ANGLES FOR SPOTTED ANTBIRDS

Foraging Away from Ants Foraging with Ants Preening Angles Records Percent Records Percent Records Percent 0- 20 ø 91 27.1 593 30.1 61 67.0 20- 40 ø 39 11.6 306 15.6 18 19.8 40- 60 ø 46 13.7 268 13.6 10 11.0 60- 80 ø 53 15.8 163 8.3 1 1.1 80-100 ø 106 31.6 628 31.9 1 1.1 100-120 ø 1 0.2 8 0.4 120-140 ø 2 0.1 Total 336 100.0 1968 100.0 91 100.0 14 ORNITHOLOGICAL MONOGRAPHS NO. 10 falls and tanglesof lianasor palm clumps. Vertical perchesare seldomused (Table 2), eventhough foraging from verticalperches is a major adaptation in this and related genera. To preen(Figure 5, a), a SpottedAntbird fluffs or elevatesfeathers of the regioninvolved, pokes the headinto that region,and with a twist of the head runseach feather from baseto tip betweenthe mandibles.As well as preening feathers,the bird nibbles at the basesof feathers. At times the bird flashesone wing far out and peers under it ("underwing-looking"). Occasionallyone shakesthe heador bodyduring a preeningsession; the latter activityis more commonduring bathing. In general,the preeningactivities resemble those of BicoloredAntbirds and other small passerines. After dissectingprey, and periodicallyduring preening or other situations, a SpottedAntbird sometimeswipes the bill energetically. It strops the bill rapidly from base to tip on the perch or a nearby surface,in much the same way as doesa BicoloredAntbird. Champingor mandibulation,opening and closingthe beak rapidly several times,is occasionalafter preening, dissecting food, and in someother situations, suchas the activelooking about after a supplanting. I recordedSpotted Antbirds scratching the headby passingthe foot under the wing only twice. Normally this and other antbirdslower one wing and scratchthe head over it. I specificallyrecorded scratching over the wing 48 times for SpottedAntbirds, and saw scratchingover the wing many other times without recordingthe events. Often the head is fluffed, especiallythe crown of the head. Once I noted that the tail was lifted as the bird scratched over the wing. Mutual grooming,when a bird groomsits mate or young,is describedunder reproductivebehavior.

STRETCHING After a SpottedAntbird rests or preensfor severalminutes, it is likely to perform a stereotypedstretch or two before returningto foragingor other activities. These stretches,common to Spottedand BicoloredAntbirds and many other birds,include yawning or the "bill-stretch,"full side-stretches(of wing, tail, and leg on one side), half-flexes(of the partly flexedwings above the back), and toe-standing.Perhaps these stretches are "negativeafterimages" of restingpostures, in that musclescompressed during restingare extended and vice versa.

EGESTION AND DRINKING Foraging behavior is consideredlater, in a discussionof the relationsof SpottedAntbirds to army antsand to other birds. At times SpottedAntbirds regurgitatefragments of insect exoskeletons, 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 15 althoughthe fragmentsnever seem to be compactedinto pellets. The excreta are white and rather fluid, and are droppedrather frequently. There are no strongbehavior patterns associated with elimination,except for slightflexion of thelegs and opening the tail covertsso the feathers and legs are not soiled. SpottedAntbirds rarely drink, exceptthat they sometimesnibble the drops of water at the tips of leavesafter rains.

SUNNING One female SpottedAntbird "sunned"for a minute. She hopped onto a sun-lit twig, half-satwith her feathersfluffed down to the perch, and spread the wing on the sidetoward the sunas sheturned her face towardthe sunon that side. Her bill was open, her crown and head somewhatfluffed. The shadedeye opened and closedas if she were sleepy,but the eye toward the sunwas opened. Ordinarily Spotted Antbirds avoid sunflecks as if photo- phobic,so sunningis probablyan uncommonactivity.

BATHING Occasionallya SpottedAntbird bathes in a shallowpool, either as a break in foragingduring the day or, more commonly,as one of the last activitiesof the day, between17:00 and 18:00. One pair chirpedand faint-sangat 17:20 as they moved slowly to a little creek, looked down at severalpools and bathedbriefly in them,then flew downto the shallowedges of two poolsabout 0.7 m apart. After hoppinginto water about to their midlegs,they sat and flutteredbriefly. Then they waited and looked about, half-sittingwith wings loose and their tails spread,before ducking their heads and forepartsand flutteringbusily. Drops cascadeddown their backsas they rose. Their tails were half in the water. The female flew up and shook her plumage,then droppedto her puddle for more dips of the body and foreparts. The two finally movedto low perchesin a nearbytreefall, where they preenedquietly as I left at 17:35. Other bathingbirds alsobroke bathinginto shortsegments by flying to perchesnear the pools briefly; perhapssome did so becauseI disturbedthem. Shakingthe body is veryfrequent after bathing but infrequent at other timesin this species.

ANTING SpottedAntbirds "ant" muchas do BicoloredAntbirds (Willis, 1967). The SpottedAntbird holdsthe prey in the tip of the beak and chewsit between sessionsof pokingthe prey into feathersunder the tail and wings. To stroke theprey underthe tail, the antbirdstands very high on the legsand lowersthe slightlyspread tail until it touchesthe legsor fills the spacebetween them. It may be that the bird is strokingthe prey on the remigesand rectrices 16 ORNITHOLOGICAL MONOGRAPHS NO. 10 themselves,but the strokingseemed to passthrough the undertail covertsor flank feathers. Out of 11 SpottedAntbirds recorded anting with 14 prey items, 5 birds ate 7 prey items. On two of the occasions,male Spotted Antbirdsanted and then fed the prey to their mates. The fate of 5 itemswas not recorded. One bird anted with 4 small prey, eating 3 of them, then preened.One item was definitelyan ant, but I was unableto identifyothers.

ATTACKS BY ARTHROPODS Spotted Antbirds occasionallyhop violently ("jitter") from one foot to the other when army ants attack their toes, but such events are rare. Many SpottedAntbirds forage away from the most active parts of an ant swarm, becauselarger birds drive them to the periphery. The ants, most numerousand aggressiveat the swarmcenter, are not likely to attackperipheral birdsvery often. I have rarely seena SpottedAntbird peck and throw away an attackingarmy ant or shakeone foot at an attackingant, but theseactions are so widespreadamong related antbirds that the SpottedAntbird probably does them whenevernecessary. Mosquitoeselicit head-shaking,flitting the wings, and twitchingthe tail from side to sidefrom SpottedAntbirds, much as in BicoloredAntbirds. I have not seena mosquitoget bloodfrom any antbird. Bird ticks occasionallyattach near the cornersof the gape and swell up with blood. Wanderingimmature birds seemmore prone to have ticks than do settledbreeding birds, suggestingthat the latter encounterfewer ticks or removethem more effectively.

REACTIONS TO RAIN SpottedAntbirds generally wait underleaves or other shelterduring heavy rains,but moveabout readily during medium to light rains. They foragelittle during mediumto heavy rains. Free-movingbirds rarely seemto have any problemswith wet feathers,even thoughthe feathersof birds capturedin mist nets during rains quickly becomesodden. A small bird like this can easilyfind local placesprotected by overheadleaves, as the rainfall pattern near the forestfloor is notablyirregular. At times a SpottedAntbird shakes its bodyor headto throw off a raindropor spray.

DISCUSSION

Such adverse(i.e., entropy-increasing)and small influencesas stiffness, desiccation,rain, cold, dishevelment,dirt and external parasitesseem to troublefree-living Spotted Antbirds relatively little in theirnormal environment in the shadedforest undergrowth. Still, maintenancebehavior against these adverseinfluences occupies some time and energy,and birds unable to spend 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 17 enoughtime, such as immatures chased about by adultbirds, quickly become disheveled. Temperaturesin theforest undergrowth (Allee, 1926: 280) are well below thebody temperatures (42.0øC; n = 34; Y. Oniki,MS) of SpottedAntbirds, especiallyat nightor duringrains. The fluffingduring resting and preening is probablya heat-conservingactivity, as is sittingso that thefeet are covered. Sittingmay alsorequire less muscular effort. The rather sparseand loose featheringof antbirds,compared with northernforest birds, noticeable when one preparesstudy skins, probably is enoughto conserveheat withoutre- quiringan inordinateamount of plumagecare or featherreplacement at the molt. I doubt that heat itselfis muchof a factor, excepta permissiveone, in the evolutionof sparsefeathering. Some tropical seabirds and birds of clearings have thick feather coats,so that the latter can protect againstradiant heat. Intensetropical heat, rare in the forestinterior, might occasionallystress the thinly-featheredantbirds if they did not avoid sunlit clearingsor open areas so assiduously.Spotted Antbirds never seemed overheated enough to show gaping,panting, gular flutter, ruffling the feathers, or sleeking.Birds released from mistnets sometimes opened the beak and pantedrapidly, as did birds persistentlychased, but thesereactions were short-livedand may have been relatedto fright rather than to overheating. Avoidanceof sunflecksand clearings, and the rarity of sunning,are probably reactionsto light rather than to heat. Gordon Orians (pers. comm.) has pointedout that antbirdsof the forestundergrowth have very large eyescom- pared with birds of the forest edge,and that they may have difficulty seeing in bright light. Timidity in well-lightedopen situationsmay be due to problemsin detectingpredators. Normal behaviorpatterns probably protect free-moving Spotted Antbirds againstheavy tropical rains rather well. Water conservationis probablynot a problemfor SpottedAntbirds, becausethey live in a constantlyhumid environmentat moderatetemperatures and eat very juicy prey. Their fluid excretasuggest that eliminatingwater and nitrogenare not problems. Preeningand restingbirds generallytake horizontalperches, but foraging ones take vertical perchesas well. Spotted Antbirds are speciallyadapted to clingto verticalperches, but it musttake lessenergy to stayon a horizontal perchif a bird has a choice. The light-seekingundergrowth of deep tropical forestsis verticallyoriented to a degreethat northernerscan scarcelyimagine; a foragingbird, especiallyone over army ants,will seldomhave a horizontal perchto standon unlessit takesthe ground,where it will rarely have a good vantagepoint and probablywould be attackedby ants. Probablyforaging SpottedAntbirds have to take verticalperches, but preeningones can hunt for horizontal ones that are easier to stand on. 18 ORNITHOLOGICAL MONOGRAPHS NO. 10

Anting is generallyconsidered a kind of maintenancebehavior, one that consistsof the use of insect secretionsand other things becausethey are stimulating(Whitaker, 1957), kill mites (Dubunin,)'Me Kelso and Nice, 1963), or sootheareas of molting feathers(Potter, 1970). In SpottedAntbirds and many other tropical speciesI have studied,anting seemsto be a way of treating objectionableprey. SpottedAntbirds sometimeseat such prey after anting, while Plain-brownWoodcreepers commonly do so. Possiblyanting may have secondaryvalue in plumagecare in both of thesespecies. Normally antbirds preen off or shake off dirt and water and external parasitesrather well. The head is a site that ticks and feather lice occupy rather frequently,however. Perhapsthe relative frequencyof groomingthe head in mutual groomingis correlated with difficulty in grooming the head. The wingis anotherplace where mites and bird lice stayin manybirds; perhaps the patternof underwing-lookingis a reactionto parasites.

REACTIONS TO DANGER

Reactionsto large predatorsare a kind of maintenancebehavior, related to reactions to small parasites,but distinctive and prevalent enough to be consideredseparately. When a possiblepredator appears, a SpottedAntbird may freeze, panic, or mob. Later reactions are fleeing, tameness,and investigating.

FREEZING

If dangeris distant or uncertain,the SpottedAntbird sometimescrouches and freezes. Related antbirds sometimes "keen" (Willis, 1967: 13) when theyfreeze, but I nevernoted this call for SpottedAntbirds. At timesa freezing Spotted Antbird gives a faint, long note, which seemsalmost a "snarl." Comparedwith the standardposture, the body feathersof a freezing Spotted Antbird are compressed.The head is up but the front of the body is down, indicatingfemoral flexing (as in Willis, 1967, "flexing" means closing the angle proximal to the part cited and "extending"means opening the angle proximal to the part cited) and head extending;the neck is retracted. The only movementsare blinking the eyes unlesssigns of panic such as head- turning and wing-flittingor tail-flickingbegin. SpottedAntbirds flushedoff nestsor travellingthrough the forest sometimes alternatefreezing with suddenmovement in an alternatingpattern that makes them difficult to follow. On one occasionfreezing was a reactionto my movement. Once it was a reaction to an outburst of chipping from Ocellated and Bicolored antbirds, once a reactionto keeningfrom OcellatedAntbirds, and once a reaction to loud alarm calls ("stieking") of Plain-brown Woodcreepers. Compared with 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 19

Figure 6. Mobbing (a) and panicking (b) female Spotted Antbirds, from field sketches.

Bicolored Antbirds, Spotted Antbirds freeze very infrequently. They more oftenstart chipping and panicking, which they do on the slightestpretext.

CHIPPING AND PANICKING At times a SpottedAntbird startschipping peep!, one or more times in each series,loudly and emphatically. Often it is more or less in a posture or display(Figure 6, b) that I call "panicking."The bird jerks its head one way and then the other. The body and head are sleeked,and the wingtips scissortogether. The bird flits the wingsbriefly as it flicks the more or less spreadtail very sharply. I am not certainwhat movementsproduce flitting; it may simplybe a brief carpalextension, but a quickflexion of humerusand midarm may contribute. The panickingbird seemshigh on its legs,but the body anglesdownward to the horizontalor below. Probably somejoints of each leg are extended and othersflexed to producethis posture;the foot and tarsusare apparently extended,the femur and tibiotarsusflexed (see definition three paragraphs above). Often the flickingand flitting lead to pivotingback and forth or reversing to mirror-imagepositions as the bird looks at the object of alarm with one eye and then the other. Thesesharp movements lead to suddenflights from perchto perchand to coverif dangerpersists or if other ant-followingbirds panicor call in alarm. The SpottedAntbird chipsloudly as it fleesto a treefall, palm clump,liana tangle,or other cover. It continuesthe displayand calling in cover. If dangeris not so close,the bird becomeshyperactive at foraging. It salliessharply for prey with peep!notes, darts from perchto perch, and in generalacts as if mixingthe activitiesof panickingand foraging. SpottedAntbirds seem very excitable,very proneto panic and chip rather than usefreezing or mobbing.While otherantbirds are chirringand mobbing the observer,the SpottedAntbird quickly turns to panickingand chipping. While otherantbirds are freezingat a distanthawk, the SpottedAntbird chips loudly. 20 ORNITHOLOGICAL MONOGRAPHS NO. 10

Bicolored and Ocellated antbirds,the dominant and well-fed speciesat swarms,sometimes react to the chippingof SpottedAntbirds and stieking of Plain-brownWoodcreepers (another subordinatebird that calls and panics readily) by keeningand freezing,chipping or panicking,or fleeingto cover. Some subordinateand poorly fed BicoloredAntbirds continueto forage as they chip and panic, as do the SpottedAntbirds and woodcreepers.Often the dominantbirds continueto hide and preen in cover rather than over the ants after the generalpanic is over, and at such times the subordinatebirds forage very busily. At times a SpottedAntbird chips when a Bicoloredor OcellatedAntbird or otherlarge bird supplantsit. I specificallyrecorded peep! calls from Spotted Antbirdssupplanted by BicoloredAntbirds on 18 occasions,by other Spotted Antbirds6 times,by Gray-headedTanagers (5), Plain-brownWoodcreepers (2), Ocellated Antbirds (2), Great Rufous Motmot (2), White-whiskered Puffbird (1), and Chestnut-backedAntbird (1). Twice a SpottedAntbird displacedby a coati (Nasua narica) gave chippingcalls, and once a fall by a nearbymarmoset (Saguinus geoffroyi) causedchipping. Chippingwas a reactionof SpottedAntbirds to hawks (Double-toothed Kites, three times; Common Black Hawk, once) and to alarm notesof other birds (Plain-brown Woodcreeperstieking, twice; the ah, rrrrrt! of a Slaty Antshrike,once). Oncethe chatterof a flushedBlack-throated Trogon caused chipping.On manyoccasions I seemedthe causeof their chippingand panicking,but chitringand mobbingwere generallytheir first reactionsto me.

CHIRRIlqG AND MOBBING I recordedchirring at a passingbrocket (Mazama americana) twice and to a runningcoati once,but the other prolongedchirring reactions were to me and other humans. Birds away from other birds are especiallylikely to chirr persistentlywhen the observerpasses. Bicolored and other antbirdschirr at passingcats, tayras (Eira barbara), and other groundpredators, but Spotted Antbirdstend to go into chippingand panickingmuch more readilyin such situations.One SpottedAntbird startedflitting and flickingwhen a squirrel (Sciurusgranatensis) approached. Others flitted and flicked and were silent when I approached. The mobbingdisplay or behavior pattern is rather similar to that of Bi- coloredAntbirds, but is even more likely to be confusinglycombined with panickingbehavior. Tail-spreadingand -flicking are probablycharacteristic of both patternsof behavior,but wing-flittingis more likely to be a sign of panickingthan of mobbing.Birds that chirr withoutadding chipping flit very little. My earlier statement(Willis, 1967: 40) that BicoloredAntbirds flick the tail more during mobbingthan during panickingis probablyincorrect; recent studies suggestthat Spotted and Bicolored antbirds flick the tail 1972 WILLIS: BEHAVIOR OF SPOTrED ANTBIRDS 21 stronglyboth in mobbingand in panicking. A SpottedAntbird combining mobbingand panickingoften turns the head one way and then the other or pivots,reverses, and darts from perchto perchor to behindcover, but mobbing birds tend to hold a perch and direct the conspicuouswhite breastand both eyes toward the observer. The mobbingbird tilts the bodyupward (Figure 6, a) and splaysthe legs, suggestingthat mobbingdiffers from panickingin extensionrather than flexion of the femora. I now rather doubt statements(Willis, 1967: 38, 41 ) that panickingBicolored Antbirds splay the legs;more likely panickingBi- colored and Spottedantbirds flex or adductthe femora and hence bring the legs toward the centerline. Probablyboth Bicolored and Spottedantbirds flex the tibiotarsiand extendthe tarsi and the feet in panickingand mobbing, but the feet seemto be flexed more in mobbing. Instead of scissoringthe wingtipstogether as in panicking,the mobbing bird generallyspreads the tips of the wings. I am not surewhether this rep- resentsactive carpal extensionor is a passivecarpal extensioncaused by humeral extension. As in BicoloredAntbirds, the neck is ordinarily flexed to the body and lifted duringmobbing, in contrastto the neck extending(lengthening) and neck loweringduring panicking. The headis often flexed on the neck during mobbing,but is extendedif panic entersthe display. I am not certain what feather movementsare associatedwith mobbingin SpottedAntbirds. Probablyhead-fluffing and throat-fluffingare used, as in mobbingin BicoloredAntbirds. The body feathersare probably somewhat spread for mobbing, as this would make the bird look bigger and more ferociousto a mammal;possibly the body is inflated,too. However, sleeking of both body and head startsas soon as a bird mixes panickingwith its mobbing. Birds with nestsor youngout of the nest chirr and mob the observervery stronglyand persistently.These birds show little panicking,and probably would be the best onesto observefor future studiesof mobbingin antbirds. In northern latitudesand brushy or densehabitats, birds often approach and call whenone makesa squeakingor hissingsound. Usuallytropical birds, especiallyones of forests,flee at suchsounds rather than approach.However, SpottedAntbirds with youngout of the nestsometimes appear and chirr when one tries squeaking,as Richard Stallcup (pers. comm.) has also noted on Barro Colorado. One male near a just-robbednest that had had nestlings in it came up and chirred repeatedlywhenever I squeaked;he gave faint "snarling" also (Figure 7). Certain individual SpottedAntbirds are especiallyprone to chirr and mob the observer. Male YBXR, who occupieda territory near the summit of 22 ORNITHOLOGICAL MONOGRAPHS NO. 10

Figure 7. Above, Spotted Antbird male mobbing or investigatingme when I squeaked near nest from which young had disappeared. Below, male ending investigation.

Barro Coloradoand saw me every year from 1960 to 1966, alwayschirred and fled in panic when I appeared. He habituatedto my following very slowly, and seemedto be as wild at each new meetingas the first time I saw him. His sonand grandsonalso chirred and mobbedvery readily. Most other 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIR-DS 23

SpottedAntbirds I have encounteredrepeatedly ignore me or chirr only briefly,even when I appearfor the first time in severalmonths.

FLEEING AND TAMENESS Quite oftenSpotted Antbirds flee anddisappear after mobbingor panicking, especiallyif they are in open undergrowthand are not with army ants or with a flock of birds. To read the bandson their legs, I sometimeshad to chase them about for half an hour to an hour before I could come within 20 m of them. They can disappearsurprisingly well behind an isolatedtree trunk or two, and changedirections so frequentlythat they can be difficult to follow. They hide in treefallsand in densepatches of saplings,behind or under palm clumps,or in tanglesof lianas, and circle in thesethickets when one tries to drive them out. Infrequently(more often if one setsa low mist net!) a persistentlychased bird flies higher and higher in the undergrowth as it doublesback and forth until it is 5 to 15 m abovethe groundlike an antwren. Poorly flying fledglingsdo this regularly. Fleeing birds are com- monlysleeked and flick the somewhatspread tails, probably as low-intensity panicking;they chip or are silentrather than chirr in manycases, unless young or nests are nearby. At times one flees until it reaches an antwren flock (see below). SpottedAntbirds often becometame. They are usuallyamong the first birds to return to a swarmof ants and resumeforaging after the observer appears, although one usually has to wait 5 to 15 minutes. Subordinate BicoloredAntbirds also return quickly, suggestingthat low statuson the peck order is directlycorrelated with tameness. While SpottedAntbirds habituate to the observerrather rapidly, they never become as tame as do individual Bicolored Antbirds. The latter habituate to the observervery slowly,but with frequentcontact become so tame that they forage within three m. PerhapsSpotted Antbirds did not becomethis tame becauseI seldomobserved an individualas long or as frequentlyas I did individualBicolored Antbirds. SpottedAntbirds are likely to be driven away from swarmsby the larger BicoloredAntbirds, or to wander about the outskirtsaway from the swarmand me. Moreover,each SpottedAntbird has a smallerhome range than doesa givenBicolored Antbird, and is encountered less often as I follow a swarm of ants about throughthe forest. However, evenSpotted Antbirds that I have observedrepeatedly over the yearstend to flick and flit or move away,if I try to observethem from closerthan five m, more often than do BicoloredAntbirds I have seenequally often. Tame birds usuallyare rather fluffed when I approach. At ames the head is fluffed, as in tame Bicolored Antbirds. The tail is closed and down, or jigglesin slightflicks. These movementsprobably are low-intensitysigns of mobbingrather than a distinct"tameness" display. 24 ORNITHOLOGICAL MONOGRAPHS NO. 10

INVESTIGATING Infrequently,Spotted Antbirds fly toward the observerafter (Figure 7, upper) or during a mobbingsession or as he watchestame birds quietly. Young birds are more likely to "investigate."The investigatingbird peers with one eye and then the other. It extendsthe head on the neck, extends the neck,and lowersthe neckon the body. The bird flits forward,suggesting that the femora are flexed. Probably some other segmentsof the legs and wings are flexed or extendedmore than in the standardposture, but the fluffing of the body hides most such movements. Moreover, I was usually uncertainwhich leg and wing movementswere thoseof mobbingor panicking and which were ones of investigating.Commonly the investigatingbird approachesand retreats(Figure 7, lower) repeatedly,suggesting ambivalent "approach-avoidance"behavior as in chaffinches,Fringilla coelebs,and other birds (Rowell, 1961: 59).

REACTIONS OF CAPTIVES In mist nets or in the hand, a SpottedAntbird kicks, openingits feet at the sametime, and immediatelyflutters its wingsor attemptsto flutter them. Sometimesit escapesand flies off. If not, it tries kick-flutteringagain and again,sometimes with a chipor two. As oneremoves it, it sometimeswhim- pers. In the hand it may start peckingor hangingon and twistingwith the sharp,hooked bill; but it is not as persistentat this as are the larger antbirds, perhapsbecause its smallbill cannothurt an enemymuch. If one freesthe tail and wings,they are spreadas the antbirdpecks. The back is sometimesspread so that the whitepatch shows. At timesthe peckingbird snarlsfaintly. There may be a brief scream,but SpottedAntbirds are not persistentscreamers. Peckingattacks may be relatedto the displayof challenging,and whimpering to the displayof cringing,displays described in the sectionon agonistic behavior.

DISCUSSION SpottedAntbirds have at least two main types of aggressivereactions to dangerfrom large animals--attackor "corneredrage" (as in the hand) and mobbing--and three main types of escapereactions--passive freezing and active panicking and escape. Two other behavior patterns,tameness and investigating,are probablyuseful when dangerof captureis slight. Someof thesepatterns are similar to ones that psychologistsand neurophysiologists have studied in rats and other mammals: stimulation of the basal nucleus of the amygdalagives rage and ablationgives tameness (Beritoff, 1965); fleeing to avoid a shock(the CAR, or conditionedavoidance response) depends on the cingulategyms, and freezing(?) to avoid a shock(CER, or conditioned emotionalresponse) depends on the septum (reviewedin McCleary and 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 25

Moore, 1965). Ablation and stimulationexperiments suggest that rage, freezing, and fleeing have connectionsupward to the cortex and downward to the hypothalamus(Brown and Hunsperger,1963), and that rage areasare roedial to escapeareas. Akerman (1966b: 341ff.) reportssimilar results for pigeons(Columba livia): medialstimulation from hypothalamicto subcortical (archistriatal) regionsgives complete sequences of defensivecalling and fighting,while more lateral sites give sequencesin which fleeing con- sistentlyfollows freezing. Mobbing and investigatingseem to be unstudied, perhapsbecause these patterns depend on complexcortical centers more than on lower ones. The observationson Spottedand other antbirdssuggest a diversityof predator-relatedbehavior patterns,which could be based on distinctneural regions like thosethat controlsome pigeon and mammalian patternsor whichcould be linked (as freezingto fleeing) in part. Many ethologistshave suggestedthat conflicting"drives" or "tendencies," such as attack and escape,are major componentsof behaviorpatterns or displays(Hinde, 1969: 381; Marler, 1956; Moynihan, 1955; Stokes,1962; Tinbergen,1959). One shouldalways question widely acceptedhypotheses, however,if there are alternatives.I questionthe hypothesisof evolutionof behaviorby meansof conflictingdrives, because I think the hypothesisof simpleritualization of unitary driveshas not beendisproven. The hypothesisof conflictingdrives, when it suggeststhe evolutionaryorigin of a display,is comparableto the theorythat a speciesarises when two other specieshybridize; the hypothesisof unitary drives is rather like the theory that speciesevolve by divergentevolution. The questionin both speciation and the evolutionof displaysis whetherhybridization or divergentevolution is more important as a source,because both patterns of evolution probably occur. Accordingto evolutionarytheories, a hybrid animal or display can perhapsoutcompete parental typesin new or intermediateor changeableen- vironments,especially if competitionor dangeris not extreme. The mobbing- panickingambivalence in Spotted Antbirds may be appropriateto a new situation,where the human approachesrather than wanderspast sucha small bird as big mammalsshould do. In normal predatorreactions, however, am- bivalentbehavior seems likely to be selectedagainst unless variable behavior confusesthe predator. The reaction to a significantpredator should be definite, quick, and appropriaterather than ambivalent. Therefore, it seems morelikely that predator-relatedbehavior patterns evolve through divergence than by hybridizationor Hegelian"thesis plus antithesisgives synthesis." The hypothesisof conflictingdrives also suggeststhe neurophysiological origin of suchdisplays as mobbing;it saystwo inputs are necessaryfor a givenoutput. Preliminaryneurophysiological work (Akerman, 1966b: 344) suggeststhat protectivebehavior comes from singlebrain areasdifferent from 26 ORNITHOLOGICAL MONOGRAPHS NO. 10

thosefor defensivethreat and aggressivethreat; there seemsto be no evidence that stimulationof two centersis necessaryto produce ordinary predator- avoidancesequences, and hence no evidence that the hypothesisof conflicting drives is correct there. Given the presentstate of knowledge,it is almost as difficult to suggest ecologicalcauses for particular predator-relatedbehavior patterns as it is to suggestphysiological causes. It is uncertain,for instance,why freezingis rare in SpottedAntbirds but moderatelyfrequent in BicoloredAntbirds. Perhaps SpottedAntbirds freeze infrequentlybecause they usually forage in open undergrowth,near but not in cover, and move about rather activelywhenever larger birds excludethem from the best and safestsites at swarmsof ants. Freezingwould be maladaptiveif the predatoris likely to have seenthe prey moving about. If so, the even more active and less frequentlyskulking ant- wrensshould show freezing less frequently than do SpottedAntbirds. Panickingand chippingare unusuallyfrequent in SpottedAntbirds. The chippingof lone birds travelingto nestsis particularlydifficult to explain. Alarm calls, accordingto usual theories,warn others of the same species, particularlymates and young. It is possible,however, that sharp and easily locatedalarm calls also tell the predatorthat "this is an alarmedbird, one that will be difficult to catch." Keening sounds,present in speciesrelated to SpottedAntbirds and in many other species,are probablydifficult to locate (Marler, 1955) and warn others of the same specieswith little risk to the caller. Chippingcalls, by contrast,are short and sharp and have plenty of cues (high frequenciesfor locationby differencesin wave form at the two ears, plus suddenor staccatoon-off for locationby time of arrival at the two ears) for localizationby any listeningpredator. I suspecta waitingpredator is lesslikely to attack a bird that chips and showspanicking behavior as it travelsthrough the forest. It may alsobe that the suddennessand irregular movementsof chippingand panickingstartle or confuseor irritate a potential predator,in the way the suddensounds and zigzagflights of CommonSnipe startiehuman hunters. If so, hyperactiveand conspicuouschipping could be selected for. The tendencyof pairs of Spotted Antbirds to forage moderate distances apart,rather than togetheras do pairsof BicoloredAntbirds, may alsoexplain why the latter useventriloquial keening moderately often in dangersituations but the former generallyuse the easilylocated chipping. The mate may need to locate the caller if the two are far apart, but keeningwould not specify the location of the caller. Foraging separatelyis often necessaryfor pairs of a subordinatespecies in poor foragingzones around swarms of ants,while pairs of a dominantspecies can get food near each other at the rich central zones. 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 27

At swarmsof army ants, the nichesof SpottedAntbirds and Plain-brown Woodcreepersmay have somethingto do with their frequent and at times seeminglyhysterical panicking. Larger dominantbirds, mainly Bicoloredand Ocellatedantbirds, take the best and safestforaging sites and leaveperipheral and unsafesites to the SpottedAntbird and the woodcreeper.Being at the edge, the first line againstpotential predators,would by itself favor quick panicking. Moreover, the dominantbirds tend to deserta swarm and let the subordinateones forage during and after a generalpanic. Otherwisethe dominantbirds tend to preempta swarmof ants by remainingover it even when preeningor resting. Selectionmight favor thosesubordinate birds that are prone to call the alarm quickly and persistently,even to suchthings as fallingleaves, if theythereby gain a little foragingtime at the centerof swarms of ants. The usefulnessof prematurepanicking would be limitedby habituation of the dominantspecies, which might ignore the alarm notes of subordinatespecies that cried "wolf" too often. Bicoloredand Ocellatedantbirds do ignoremany of theisolated peep calls of SpottedAntbirds; it takesa nearby or repeatedseries from the latter to start the first two in their own chipping panics,although they look up or freezebriefly at nearly every chip of a Spotted Antbird. There are thus at leastfour likely functionsof hyperactivepanicking and callingin SpottedAntbirds: warning or teachingothers of thespecies; warning a predator;confusing it; and scaringoff dominantcompetitors. Mobbingbehavior probably has similarfunctions in Spottedand Bicolored antbirds:irritating or distractingthe predatoryanimal that is near a foraging site or offspringso that it movesaway; and warningor teachingother birds of the species:I have elsewheresuggested that chirringmimics the growls of carnivores,and that suchharsh bird calls are unpleasantto humansbecause our ancestorshad to live and hunt amongbirds that couldscare away our prey (Willis, 1967:41 ). Chirringis relativelyeasy to locate,even if the high frequenciesthat allow locationby differencesin intensityat the two ears are lacking. Mobbing behavior seemsto be threateningrather than escape behavior, even when the birds dart to cover or hide behind tree trunks or other vegetation. It is generallydirected at predatorsthat are not likely to catch an alarmedadult, such as slow-movingmammals. Rarely other speciesof antbirds(Willis, 1969a: 369) chirr at slow-movingor waitinghawks; probably chirring is not a "ground-predator"call but a general "disturbance"call. Hawks are generallymore dangerousthan simplydisturbing, of course,and antbirdsgenerally react to them by chippingand panickingrather than by mobbing and chirring. Thereare several possible causes for therelative lack of reactionto squeaking noisesin SpottedAntbirds and other birds of the undergrowthof tropical 28 ORNITHOLOGICAL MONOGRAPHS NO. 10 forests. Probably it would be dangerousfor a slow-flyingbird of the open undergrowthto approach such noises, for forest-falconsand other hawks sometimesapproach such noises (N. G. Smith [1969] and Dennis Paulson, pers. comm.). Approachto squeakingseems characteristic of birds of dense cover, birds that have plenty of placesto escape. These birds are also more likely to gain by approaching,for they can seldomsee predationor learn predatorsfrom a distanceas can birds of more open habitats. It may also be that birds of bushyor northernhabitats, with their larger broods (Lack, 1968: 166-168), tend to mob or approachnoises that soundlike their own young; such actionsmight distract a predator that has caught one young, thus protecting others. Robert Ricklefs (pets. comm.) has found that tropicalbirds of specieswith broodsof four defendtheir youngmore vigorously than do ones of specieswith broods of three or two. Experimental studies, plus comparativeanalyses of responsesto squeakingin differentbirds and age classesat different seasons,are needed to determine whether such re- sponsesserve to protectyoung, allow learning, or performsome other function. Other kinds of curiosityalso seem more characteristicof birds of dense habitatsthan of tropical birds of the forest interior. SpottedAntbirds are relatively incurious,tending to flee rather than approachin almost every situation. Only young approachreadily. Possiblybirds of densevegetation can afford to investigatean intruderwith little danger,given that investigating has some advantage.Hermit Hummingbirds,especially the tiny speciesof denseundergrowth ( Phaethornisruber, Phaethornislonguemareus ) investigate humansreadily inside the forest;they are able to fly away quite rapidly, of course. It may be that taking a closelook allowsa bird to seeif an animal is a predatorby allowingit to make an attemptin a situationwhere the bird has a very goodchance to escape. Probably the relatively constantmoderate tamenessof Spotted Antbirds, as contrasted with the initial wildness and later lack of fear of Bicolored Antbirds, is related to various features of their ecologies. The Spotted Antbirdis small, and probablyis lesslikely to be prey for a large mammal than is the Bicolored Antbird. The even larger Ocellated Antbird is much more wary than is the BicoloredAntbird. However, elsewherein the class Aves large birds are not alwaysmore timid than small ones, so that there may be other reasonsfor differencesin tamenessin the three antbirds. The SpottedAntbird nestsabove the groundrather than in holesin stumps like the BicoloredAntbird, henceis lesslikely to be trappedin a nest by a mammalor have a mammalrob its nest. The SpottedAntbird is low on the peck order when larger birds are arounda swarmof ants,but can get prey if it foragesnear the observerwhen wary larger birds retreat. (Subordinate individual Bicolored Antbirds, and the similarly subordinateLunulated 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 29

Antbirds of the upper Amazon Willis, 1968: 137--also forage near the observer and become tame readily.) However, the Spotted Antbird is a facultativeant-follower and can wander widely around a swarm, so that it need not stay near the observerand become extremely tame the way subordinate Bicolored Antbirds must do. SpottedAntbirds also must forage in rather open siteswhen large antbirdsexclude them from sitesnear cover, hencemust be very alert and wary.

AGONISTIC BEHAVIOR Agonistic behavior, or the reactions animals show when disputing over spaceor food or someother resource,includes in SpottedAntbirds submissive displays,escape behavior, aggressive displays, attack behavior,and fighting. In agonisticbehavior, as I define the term, the "enemies" are intermediate in size and distancebetween close but very small or dispersedones (e.g., ectoparasites)eliciting maintenancebehavior and distant but large or very dangerousenemies causing alarm behavior. I realize that some ethologists include a much wider range of displays,such as mobbingbehavior, fleeing from predators,etc., within their definitionsof agonism. However, I prefer to includethese activities under "Reactionsto Danger,"above.

SUBMISSIVE AND ESCAPE BEHAVIOR

Submissivebehavior is rare in Spotted Antbirds. The most characteristic movementof the display, which may be called "cringing" becauseit is similar to that displayin BicoloredAntbirds (Willis, 1967: 47), is ruffling the feathersof the head. The bodyfeathers are oftensleeked, but at timesthe subordinatebird fluffs the ventral feathers. Generally the white area on the breastis small, almostconcealed despite the fluffing. The tail is closed and droopsweakly below the normal for the standardposture. The bird may crouch,or flex the legsso that the body is closeto the perch. At times the headis downand retractedsomewhat, unless the bird is foraging. At the instant the dominant bird supplantsit, the submissivebird may give a weak whimperingcall. Sometimesthe subordinatewhimpers each time it changesperches, even thoughnot attacked. Wanderingindependent immatureswhimper more readily than do youngbirds or adults,which gen- erallyflee withoutcalling or flee givingaggressive displays or displaysrather like panicking:One immaturefemale fluttered the tips of her wingsas she whimpered. Escapebehavior is well developedin SpottedAntbirds. As a subordinate bird wandersabout in a more or less submissiveposture, it looks upward or extendsthe neck upward, peeringabout as if foragingactively. One bird extendedthe headup threetimes as if peckingatop an imaginaryleaf (Figure 30 ORNITHOLOGICAL MONOGRAPHS NO. 10

Figure 8. SubordinateSpotted Antbirds, from field sketches. a, First-year male repeatedly pecks in the air over his head when chased by resident male. b, First-year female takeshead-down-and-fluffed posture after pursuitby residentfemale.

8, a). A subordinatebird changesperches very often. It tendsto moveup- ward or peripherallyif attacked,much as any SpottedAntbird doesif larger birds exclude it from a swarm of ants. At times subordinate birds moved over and foragednear me as soonas dominantbirds attacked. They may flit the wingsand flick the tail, two signsof high but unspecificarousal or excitement.A subordinatebird flips or pivotsone way and then the other, or reverseson its perch. Usually it keepsits back toward a dominantbird. As the dominantbird dartspast in a supplantingattack (see under "Wan- deringYoung and TerritorialAdults," below), the subordinatebird quickly yawsor pitchesor fluttersoff to the side. If attackedin flight, it evadesthe directcourse of thedominant bird by droppingto a perchor flutteringto one sideor the other. The subordinatebird fluttersrather conspicuously on the tipsof its wingsin flight,suggesting that thereis somerestriction of flight by flexionof the inner segmentsof the wing, as in "cringe-flying"in Bicolored Antbirds (Willis, 1967: 49). However, I have not noted suchextreme mix- tures of submissivebehavior and flight for subordinateSpotted Antbirds, though they often seem rather fluffed and have retracted necks and feet in flight. The whimperingcall is oftengiven in flight,at themoment the dominant bird darts past.

AGGRESSIVEDISPLAYS AND ATTACK BEHAVIOR The basic aggressivedisplay may be called "challenging,"because it is very similar to the displayof challengingin BicoloredAntbirds. However, SpottedAntbirds usually exhibit challengingat two "typicalintensities" (Morris, 1957) ratherthan performingmany intermediate displays. One kind of challengingis usuallyassociated with bugling,the other kind with 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 31

.. Figure 9. Challenging and fighting among Spotted Antbirds, from field sketches.a, Male in high-challengingposture. b, Male low-challengesat other male below and to his left. c, Two males flutter downwardpecking at each other. d, Typical low-challenging posture, showing forward movement of head and downward display of white chest (arrows) and a cross-sectionof the body (at right, below) to indicate fluffing of the belly feathersand parting of back feathers. e, Male in extremely fluffed low~challenging pose, confronting another challenging male. snarling.The first kind of challengingis probablyan extremeor intenseform of the second kind, in which muscles contracted for the second kind are contractedmore strongly.It may be called"high-challenging," in contrastto "low-challenging"for the displayassociated with snarling. The two kinds intergrade,but intermediatedisplays are not common. High-challengingis the first and very evanescentresponse to a new intruder. It involves,besides bugling, a suddenjerk into an uprightpose (Figure9, a). The "rule of angles"(Willis, 1967: 54), that anglesbetween the longitudinalaxis of the bodyand its limbs,head, feathers, etc., are opened for proximalparts and closedfor peripheralparts in aggressivedisplay (and vice versafor submissivedisplay) is followedin nearly every movementof high-challenging.In other words,the ends of the limbs and the head (with bill closed) tend to be flexed, while the inner portionsof the limbs and the posteriorneck are extendedfor aggressivedisplay. The head is sleekedand the body feathersexpanded, so that the white basesof the back feathersare spreadinto a conspicuouswhite patch. Probablythe bird inflatesthe body as well as fluffsthe feathers.The tall is widelyspread, and jerkedto the line of the body. I havenever seen a SpottedAntbird spreadits wingsfor high- 32 ORNITHOLOGICAL MONOGRAPHS NO. 10 challenging,but the femora and other joints of the legsare extended. At the sametime, the phalanges are flexed and grasp the perch strongly. As the neck is extended,the head is flexed to bring it to horizontal. The eyes are ap- parentlyconverged as the bird staresat the opponentbinocularly, but the movementis not as noticeableas in the larger and bare-facedBicolored Antbird. High-challengingseldom lasts more than the first secondthe challenger seesan intruder. Buglingis rarely repeatedeven oncebefore the bird drops down into a very spreadlow-challenging pose and snarlsrepeatedly at the other. At timeshigh-challenging is repeated briefly as punctuationin a long sequenceof low-challenges,but low-challengingis the stapleaggressive display of Spotted Antbirds. Basically,low-challenging (Figure 9, b, d, e) presentsto the opponentthe white back patchand whitebreast, outlined below by the necklaceof spots that givesthe speciesits name. Low-challenginggrades at lowestintensity into the standardposture (but white breastand dark tail are slightlyspread) and at highestintensities approaches the high-challengingposture. It is thusa gradeddisplay, one that endsat timesin posturespassed through on the way to higher-intensitypostures at other times. At normally high intensitiesof low-challenging,the whole chest seems to expandor swaytoward the opponentas the bird snarls.In part this is a resultof actualexpansion of the pectoralregion during or for soundpro- duction,in part a resultof fluffingof the chestfeathers, and in part a result of a forwardand downwardswaying of the whole body from the legs. The head staysin placeor is extendedforward in a slow-motionjab as the body first swaysforward and from side to side with the snarl, but the pectoral regionof the bodygoes down as if pushedfrom above. The depressingof the pectoralregion, besides expanding the white chest and dark necklace,presents the white back patchto the opponentas a frame for the dark head,which is sleeked so that it blocks view of the back rather less than would otherwise be the case. Sometimesthe head movesto one side during the snarl, thus showingthe white back even more conspicuously.The bird may reverse severaltimes, alternating showing the whitebreast and backto the opponent. At timesthe snarlingbird pivotson the perch,or sways60 to 90 degrees,as if wavingthe whitebreast and back-spot at the opponent.If the opponentis abovethe low-challengingbird, it anglesthe body upwardso the spotsstill point at the opponent. The ventralfeathers, including the belly feathersbut not the throat feathers, are fluffed out so that they completelycover the lower edgesof the wings (Figure 9, d). However, the white spotson the shouldersof the male remain exposeddespite dorsal and ventral fluffing. The tail is spreadwidely and 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 33 raised to the line of the body, or to horizontal. Betweensnarls the tail is flicked, or rather droppedand raised very slowly and regularly. If the tail is down at the start of a snarl, the bird raises it to horizontal very slowly rather than flicking it up. At times the tail is flicked as fast as once a second,but suchbirds generallystart chippingloudly as if goinginto the display of panicking. There is some extensionof the neck and archirig (flexing) of the head on the neck, but not as much as in high-challenging.Since the body angles downwardor is level for most low-challenging,some of the joints of the legs mustbe flexed. Possiblyit is the foot that is flexed to bring the bird swaying forward and downward,but it may be that more proximal joints are flexed in contradictionto the "rule of angles." The body feathersare often so fluffed the legs cannot be seen,but the bird is so close to the perch there must be either little extensionof the proximal segmentsof the legs or someflexion. The wings are sometimescrossed at the tips, suggestingeither carpal flexion accordingto the rule of anglesor more proximal flexion in contradictionto the rule. The low-challengingdisplay shows a conspicuouspatch on the back and one on the chestto the opponentat the sametime. The GuiananRufous-throated Antbird shows a patch on the back to the opponentby using a similar low-challengingdisplay (Willis, 1967: 50-52). However, in the case of the Rufous-throatedAntbird the legs are splayedand at least the central joints of the legs tend to be slightly extendedduring low challenging. Whether the level positionof its body is causedby femoral flexing or by foot flexingis uncertain;if causedby femoralflexing, the rule of anglesis con- tradictedfor a speciesvery closeto the BicoloredAntbirds for which it was first proposed. AttackingSpotted Antbirds generallydarted past the opponentso rapidly I couldnot seeany signof "challenge-flying"as in BicoloredAntbirds. How- ever, the attackersometimes gave a twit! as it flew at the other, suggestingthat challengingmay be mixed with flying to the attack. Normally the attacker givesa hissor hiss and snap as it passesthe trespasser.Beforehand, the attacker commonlycompresses the body to closeto the standardpose, or pivots one way and then the other as it staresbinocularly at the opponent. There is also commonlya pausein the snarlingand displayingfor severalseconds after an attack. Probably the attack behavior tends to replace challenging displayin mostcases, rather than the two mixing. Often the attackingSpotted Antbird flies long distances,up to 10 or 20 m, in supplantingan opponent. Since the attacker goes past the opponent severalm in many cases,the opponentoften just hops a few cm at the attack. The attacker may then do severalmore long supplantingflights, forcing the 34 ORNITHOLOGICAL MONOGRAPHS NO. 10 opponentto skipabout and stop foraging or movefarther and farther from the ants. The attackingflight seemsmore a displaythan a functionalsupplanting in thesecases. However, stroboscopicphotography will be necessaryto see whetherthe attackeractually shows certain movements of challengingas well as normal flight movements.

SEQUENCES OF AGONISTIC BEHAVIOR A moment after a trespassingnonterritorial bird appearsat a swarm or elsewhere,a resident bird high-challengesbriefly and then low-challenges repeatedly.At the first snarls,the trespasserruffs its head,flits and flicks,and foragesbusily as it circlesevasively. At times the residentjerks into a high- challengingpose without calling, or darts over and supplantsthe trespasser onceor twice. Then it followsthe trespasserabout with snarlsand occasional supplantings.If the birds are at a swarm of ants, the trespassermay return repeatedlybut is driven off each time it returns; it has to forage in poorer peripheral zones. Its feathers become disarranged,for it has little chance to preen. Often I encountereda trespasserwandering forlornly back near the ant bivouacor up to 100 m from the swarmitself. Such birds sometimes movedin as I arrived and foragedactively for a few minutesduring the com- motion causedby my arrival, but were expelledwhen the residentsreturned. Sometimessubordinate birds open the white back and go into weak low- challengingposes, with or without snarling,before they go into submissive displays or escape behavior. After a time a residentbird supplantsa trespasserwithout snarlingor challenging,but even in the nonbreedingmonths of Decemberto April there is seldomas much toleranceof a trespasseras in BicoloredAntbirds. Usually the trespassermust move out of sightof the residentbird to avoid occasional supplantingsfrom distancesof 5 to 20 m. However, quiet trespassersare not pursuedfar, and sometimesforage at different forks of an ant swarm or behind cover. Feudingterritorial Spotted Antbird malescommonly open a sequenceof agonisticdisplays with reciprocalhigh-challenges, then drop into low-challenges and snarl at each other repeatedly. At times a bird jerks up into high- challenges,with or without bugling notes, between low-challenges.If two or more territorial Spotted Antbirds feud and neither gives way, white- backed and challengingbirds swirl about like snowflakes. Some of the feuding birds repeat their snarls so intensely and rapidly and puff out so stronglythat one expectsthem to explodeat any moment. A bird may snarl as rapidly as once per two or three seconds,although the usual rate is one every five secondsor so. One bird hissesas it supplantsanother, both snarl backand forth, and the first bird supplantsthe secondagain. (Johnson[1954: 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 35

56] mistook white-backedfeuding for "courting" behavior.) Submissive behavior,especially whimpering as the victor attacks, sometimesoccurs briefly as the defeatedbird flees. Outbreaksof chippingor songssometimes end the dispute. Usuallyone femaledisputes with anotherand one male with anotherin feuds, but at times there are supplantingsbetween the sexes. Once an excited male supplantedhis own mate during a dispute. Females generally display and attack each other less stronglythan do males of the same or similarage classes, but otherwisedispute and singacross territorial boundaries much as do males. Fights and reversesupplantings are rare, even in the most heated and prolongeddisputes between rival SpottedAntbirds. I saw three brief scuffles whenbirds tried to supplanttrespassers and the latter foughtback. Once two males did a flutteringfight down from four m above the ground,pecking at each other but not touching (Figure 9, c). On another occasiona fight on the groundended in a dominancereversal, the male from one side gaining dominanceat a spot near his territory and thus incorporatingit into his area. Normally birds at or near territorial boundarieschallenge each other briefly or flit their wings, flick their spreadtails, and drift apart to sing back and forth. The strongestdisplays and attackscome from a male well within his territory, especiallyif a trespassergives aggressiveor threat displaysover a swarm of ants. Agonisticdisplays seldom last long away from swarmsor after one pair deserts.However, pairs or individualssometimes sing back and forth very loudlyfor severalminutes. Distant songs of a trespasseroften startthe resident male movingtoward it. In one casewhere I was very closeto a resident pair when a distant bird sang, I noted that the resident male mixed twit! notes with his chirpingto his mate, but did not go into the high-challenging posethat normally accompaniesthe note. Both of the pair movedtoward the other bird, and a feud was soongoing. Young birds that singquickly attract adult birds, which drive them off. When I use recordedsongs to look for pairs, the white-backedmale often comesup rapidly and silently,spread tail flicking, and looks about;he starts loud-singingafter the tape recordingstops. If the playback continues,he flies over the recorder and looks down with faint snarls but stays in weak low-challengingposes rather than goinginto strongdisplay. He pivotsone way and then the other, flits the wings and flicks the tail, darts from one perch to another,and in generalseems agitated although wary of both the observerand the recorder. The female may appear and singif he startsloud songs,but sheis generallyless excited and responsive.Recorded songs never attracted a female when her mate was absent. 36 ORNITHOLOGICAL MONOGRAPHS NO. I0

During prolongeddisputes there are often activitiesother than simple agonisticbehavior. Commonly,if the observeris close,one or more birds mayburst into chirringeven if theyhad beenignoring him beforethe dispute. At othertimes, especially in prolongeddisputes, one or more birdsburst into persistentchipping, flitting, and flicking as if panicking.At timesa dominant bird went into chippinghysteria when a subordinatebird stayedinstead of fleeing. Suddenattacks on and poundingof insectsby the antbirdsseemed moreviolent and conspicuouswhen theseactivities interrupted disputes than when they occurredwhile the birdswere foraginguninterruptedly. Once a dominantmale regurgitated insect fragments before he attackeda subordinate bird. A dominant female did the same on another occasion. One dominant male scratchedover the wing after an outburstupon the arrival of a new female.The latterstopped chipping and started chirping to him. Subordinatebirds often billwipe,flit the wingsand flick the closedtail, pointthe bill towardthe toeswhile rotating the headon the neckso oneeye is towardthe otherbird ("toe-looking"),and performother activities. One dominantbird, white-backed,approached an opponentbut then closedthe back patchand billwiped;a minutelater the first bird startedchallenging displays.Another dominant male, faced with a subordinatethat wasbusily eatingprey and ignoring him (exceptfor chippingnotes) went into a low- challengingdisplay and gave a faint bugle,then startedchipping and reversing,looked down at the trespasser,and flew aheadto the ants as the trespasserretreated. Energetic preening sessions often interruptdisputes. Foragingoften interrupts prolonged disputes over swarms of ants,sometimes in a periodicmanner that suggests the bout structure of manyanimal activities (Marler andHamilton, 1966: 153). The dominantbird grabsan insectand flailsit energetically,while the otherbird movesoff and foragesfor a time. A periodof silentforaging ends in anothersnarling and supplanting outburst. Periodsof foraging,challenging outbursts, and chases may follow each other all morningif the trespasserdoes not desert.

DISCUSSION For over a decade,Tinbergen and other ethologistshave suggestedthat threat and appeasementbehavior (i.e., aggressiveand submissivedisplays), as contrastedto fightingand fleeingand attacking,arise mainly when the animalfaces stimuli that evokeattack and escapebehavior at the sametime. For support,Tinbergen (1959: 34) notesthat a gullon its territorycommonly attackswithout display and the trespasserflees without display,while two gullsat a boundarycommonly display strongly. Neither Spotted Antbirds nor BicoloredAntbirds (Willis, 1967: 97) show this kind of behavior. At the boundary,there is very little cringingor challengingdisplay and muchdis- 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 37

placementbehavior in BicoloredAntbirds. Song disputesare more common at the boundaryin SpottedAntbirds. Challengingoccurs at boundariesin SpottedAntbirds but is strongestwhen the trespasseris well insidethe territory, and cringingis strongestfrom the trespasserwhen it staysrather than fleeing. One of Tinbergen's students,Blurton Jones (1968: 147), found that captive Great Tits (Parus major) do react with some types of threat to alarming and attack-provokingstimuli presentedat the same time. This favors the theory that conflictingtendencies of attack and escapeproduce threat. However, Blurton Jones' birds also threatenedto attack-provoking stimuli alone, especiallyif distractingfood in anothercorner of the cage or wire screensprevented simple attack. He proposesthat anythingthat interfereswith the attackof an attack-provokingstimulus may elicit threat to it. Escape tendencyis thus not necessaryto threat, even if it is one of several thingsthat caninterfere with attackand makemore advantageousthe alternate behavior of threat. Neurophysiologicalevidence also raises doubts as to the necessityof escape tendenciesin threat behavior. Von Holst and yon St. Paul (1963) and Akerman(1966a) foundthat single-electrodestimulation of "threat"sites in chickenand pigeonbrains inducedattacking and threat behavior. Con- current stimulation of sites that induce escape behavior interfered with threat. Possiblythe neurophysiologistsfound only output pathwaysof attack and escaperather than the input pathwayssuggested by Tinbergen'stheory. However, their evidence taken at face value suggeststhat threat is part of attack pathwaysand that escapeinterferes with it. Whether threat is a substitutefor attack as Blurton Jones suggests,or whetherit is part of attack as the neurophysiologicalevidence suggests, one no longer has to acceptTinbergen's early theory that escapemust balance attack to give threat. Threat need not occur only at places where attack tendenciesequal escapetendencies, as at the boundarybetween territories; it can occuranywhere that it is more advantageousthan simpleattack would be. This allowsone to seeTinbergen's observations on gulls and mine on antbirds as complementaryrather than discordant. Perhapsgulls attack in the center of their territoriesbecause they have smalland safeterritories from whichthey can chaseintruders with little effort or danger. Antbirds might threaten vigorouslyand attack for short distancesbecause chasing for long distances consumestime and energyin a large territoryand is dangerousin the tropical forest,yet a challengernear the centerof a large territory or a swarmof ants couldusurp much of the territoryor swarm. At the boundaries,gulls perhaps threaten becausesmall changesare relatively important in a small territory but attacksare preventedby dangerouslystrong resistance. Bicolored Antbirds 38 ORNITHOLOGICAL MONOGRAPHS NO. 10 do not threaten,and SpottedAntbirds only threaten a tittle before briefly singingback andforth, becausesmall changes in a largeterritory would not be enoughto offsetthe risks of predationand loss of time andenergy involved in boundarydisputes. Moreover, gulls set up territoriesrapidly while antbirds hold territoriesfor long periods;the latter situationprobably requires less threatper unit time and reducesthe rewardsof persistentagonistic behavior. The gull and antbird studiesadd to experimentalevidence that threat is an activitythat can be usedwhere it is ecologicallyadvantageous, rather than one that is necessarilylinked to boundariesor to other placeswhere escape tendencies balance attack tendencies. Someof the behaviorpatterns of SpottedAntbirds at boundariesand in intervalsbetween disputes, especially preening and billwipingand perhaps evenunusually conspicuous foraging, fall in the classof "displacementactivities." These activitiescommonly arise when a bird is changingfrom one activityto another,and henceseem more likely to be due to "conflicts"of two or more stimulior tendenciesthan doesnormal agonisticbehavior. There is currentlymuch discussionwhether such activitiesare actually relevantin situationswhere they seemdilemma-generated or out of context. Many ethologistsfeel that somedisplays were originally displacement activities; if so,there should be displacementactivities that are in the processof becoming displays.Evolutionarily, it seemsunlikely that displacementactivities will be favoredunless they becomerelevant to communicationor to someother function. Kruijt (1964) and otherssuggest that they may actuallybe relevant in suchagonistic encounters, in the sensethat differentdisplacement activities maybe performedat differentfrequencies by subordinateand dominantbirds and thuscommunicate agonistic arousal or status. Smith (1966:16) suggests that in flycatcherscertain activitiesand calls communicatetendency to take wing or not to take wing and thus are relevanteven thoughthey often seem out of place. For SpottedAntbirds, I have the generalimpression that dominantbirds are more likely to go into chippingand hystericalflitting-flicking or into mobbing,and subordinateones to perform billwiping,preening, toe-looking, andthe like. Theseactivities could be part of the generalspectrum of agonistic displaysor on the way to incorporationin that spectrum.I have described most of theseactivities elsewhere because they are primarily used in other contexts.Toe-looking, the only exception,is a very infrequentand indistinct patternin SpottedAntbirds. It is betterdeveloped in speciesof antbirdsthat displaybright-colored bare facesby it, oftenas an "eyespot-display"reaction to the observerin mobbingor investigating(Willis, 1969b: 378). Tinbergen's original theory (1940) that displacementactivities arise from high arousalor a centralexcitatory state, combined with van Iersel and Bol's ( 1958: 85) theorythat two effectivelyequal activitiesmutually inhibit 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 39

eachother and permit the appearance of a third(the "disinhibition"hypothesis),may help considerably in explaining the hysterical chipping that appears whentwo SpottedAntbirds terminate a vigorousdispute and moveapart or resumefeeding. As a fightends, a combatantpresumably changes from attackmotivation to someother motivation, such as flight or feeding;and at the momentwhen the old andnew motivations are equala thirdmotivation, suchas for chipping,may temporarily be released,or disinhibited,if the bird is excitedenough or hasthe third motivation anyway.

REPRODUCTIVE BEHAVIOR Courtshipbehavior, behavior used in pairformation or in maintainingthe pairbond between mates, involves four main activities in SpottedAntbirds: wanderingand loud-singing; chirping and "flirting"; courtship feeding; and mutualgrooming. Pair formationwill be discussedafter the descriptionof these main activities.

SINGING AND WANDERING A maleor femalethat has lost its mate, or a youngmale that has success- fullysettled in an areaby challengingand singing at neighboringmales until theyrecognize his claim,sings rather loudly and frequentlyas it travels aboutits area. Young females and males, by contrast, wander without singing. A maleor femalealso sings loudly if onecaptures its matein a mistnet or if themate disappears for sometime. If thebirds of thepair are incubating or caringfor young,there is almostnever such behavior. Infrequently, but more oftenthan do matedbirds, an unmatedbird wanders far off its territory(in Figure19, for instance,male YBXR wanderedfar fromhis territory only during1965, soon after he had lost his mate), but at suchtimes it singslittle. Neighboringbirds react with loud-singing, and song or challengingduels may sendthe wanderer back to its own area.

CHIRPING AND FLIRTING The first reactionof a lonemale to a newfemale or viceversa is usually a bugleor a snarland brief challenging, with white back showing. Then the unmatedmale or femalestarts chirping and ruffing the headas it movesnear the newcomer,rather than chasingit off. Soonthe malealternates series of faint-songswith his chirpsas "serpentine-singing,"and the femaleanswers withchirps. The body seems somewhat fluffed and the legs splayed. Together, chirpingor serpentine-singingandthe new posture may be calledthe display of "flirting." Flirtingand foraging alternate as the unmatedbird staysnear the new oneor wandersnear it. Flirtingrarely seems to interferewith foraging. Brief 40 ORNITHOLOGICAL MONOGRAPHS NO. 10

Figure 10. Courtship feedings by Spotted Antbirds, from field sketches. a, Male feeds female, who squats and raises tail as she gapes. b, Female chewing prey after feeding, in crouchedposture with neck rather extendedand tail up. c, Female chewing prey, her tail down.

snarlsand other signsof agonisticbehavior quickly disappear. The male ratheroften does "displacement activities," such at toe-lookingand billwiping, but noneseem ritualized. Ratherfrequent flicking the tail and flittingthe wings alsosuggest generalized excitement rather than definitecourtship displays.

COURTSHIP FEEDING As in Bicoloredand other antbirdsI have studied,the male SpottedAntbird feedsthe female. After chewingand preparingprey, he startsrapid chirping and serpentine-singing.The male wandersabout, chewingthe prey. When the femaleanswers with a brief faint-songor chirpingor he seesher, he flies toward her. If the female is distant, she singsloudly in answer to his loud- songs. In one case a singlesong from the female was enoughto bring the male to her throughthe forestfrom at least 50 m away. As he flutters up and feedsher, his serpentine-songsrun into a very rapid seriesof chirpsat about five per second.Her chirps,if any, comeat two per secondor less. At times the watchingfemaie darts to the male as soon as he starts chirping or serpentine-singingwith food in his beak. One female looked at her mate as he capturedprey, then looked at the rocks below her again; only when he gavethree chirpsdid shedart to him for the feeding. Both birds tend to extend the legs for a feeding,but the toes are not clamped. The legs are splayed. The female gapesas she stretchestoward the male. During the feeding,the male has his head somewhatruffed, but his 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 41 body is sleeked. The female ruffs her head as she takes the food, but she is rather sleeked(Figure 10). Her tail rises to the horizontal and is slightly spread.Both birdsmay flit the wingsif the feedingis slowor difficult. Both birds stretchtheir necksin passingthe food, or the female may grab it as the male recoilsslightly. There are sometimessigns of hesitancyor agonisticbehavior during a feeding. One unmatedmale snarledand challengedthe mate of a neighboring male while that male was elsewherewith a fledgling. The unmated male tried to feedher twice,but his attackswith white back patch showingat the lastmoments drove her awayboth times. In his abortiveattempts to feed her, he gavechirping notes that gradedto chippingones, and madeagitated flitting and flicking and back-and-forthmovements. Once a male pecked at a femaleafter feedingher, and shefluttered away. One femalepecked so rapidlythat the male flutteredaway; a second,less violent peck gainedher thefood. On oneoccasion, mates had a tug-of-warback and forth beforethe male let the femaletake the food. On anotherthe femaletried to cling near him and peckit from his beak;he flew off 2 m and shefluffed and sat down; when he fluttered off and serpentine-sangshe merely sat and gave a few faint-songs.On still anotheroccasion the femalehopped repeatedly toward her mate,who sidesteppedeach time; he wasgiving some faint snarlswith his chirps,perhaps because I was near. At times one or both birds have the white back showing. Normally the femalecrouches, head down, and eatsthe prey as the male fluttersaway after a feeding. She may hop away from him beforeeating, especiallyif shesidled to him and stretchedhis way ratherthan allowinghis approachto within one or two body widths. She may flutter off, her crest still raised,and wipe her beak after the feeding. In one caseshe froze, head low and extendedon the neck,as he flitted besideher; sheate the prey only whenhe left, thenlifted her headand fluffedher bodyto the normalresting pose. After a feeding,the male sometimessits near the female or nibblesat her beak. Groomingfollowed such feeding-sitting on someoccasions. The two birdssometimes stand near eachother for a moment,facing in the sameor differentdirections. Both or one may champthe beak or flit the wingsand flick the tail. Frequentcourtship feeding is the rule as a male gainsa new mate and in the daysbefore a nesting.If thereare no competitorsat a swarmof ants,the malefeeds the femaleso frequently that shescarcely forages for herself.She restsand preens in covernear the swarmwhile he worksactively. At timesthe femaleis so well fed that shewill scarcelytake food. In one suchcase, a malechirped rapidly as he flewup to theresting female, but shesat and looked at him. Finallyshe accepted the food, a cricket,lackadaisically as he fluttered 42 ORNITHOLOGICAL MONOGRAPHS NO. I0

Figure 1 I. Mutual grooming by Spotted Antbirds, from field sketches. a, Female stretches to preen male, who tilts and ruffs his head. b, Female nibbles bands of male, but he still ruffs the head. c, Female preening her female offspring, who tilts the head very strongly. the tips of his wingsand chirpedrapidly. As he returnedto the ants, she shookoff and ate a leg of the cricket,but the body fell to the ground;she lookeddown at it but did not go down and pick it up. On anotheroccasion, a female droppedthe prey four times and the male flew down, got it, and gaveit to her again until she finally ate it on the fifth attempt. However, thereare usuallycompeting larger antbirds at swarmsof ants,so that the female SpottedAntbird ordinarilyhas to foragefor herselfmore than do femaleBi- coloredor OcellatedAntbirds similarly close to nesting. I have never seencourtship feeding after incubationstarts or before two fledglingsare startingto feedthemselves, at about30 daysout of the nest. If thereis onlyone fledgling, cared for by the female,the male occasionallyfeeds his mate soonafter the fledglingleaves (see below).

GROOMING Flirting or courtship-feedingbirds, especially ones mated a long time, oc- casionallycome up to eachother and startmutual grooming or "allopreening" (Cullen, 1963) whenthey meet in preeningor nestingsituations rather than foragingsituations. Typically one bird hopsor fluttersup besidethe other, which immediatelyruffs the head and archesit on the neck (i.e., raisesthe neck on the body, flexes the joints of the neck or retracts the head, and flexesthe head on the neck) so the bill pointssomewhat downward. In this 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 43

"invitation-to-grooming"posture the head may be angledaway from the mate, which beginsto peck and nibble at its bill or at feathersof the crown or face and throat as if removingand eatingtiny particles. The headis normallythe only part of the bodygroomed. Once a femalepecked at the undertailcoverts of her mate, who flew off. Bandson the legssometimes get pecks(Figure 11, b). Oftenone or bothbirds champ the beakbusily between nibbling sessions. Rather frozen stancesare characteristicof birds being groomed. Com- monly the bird being groomedis sitting or sits more closelyeach time the groomernibbles it. The groomedbird may sit with tail and head down, like a towel frozen over a clothesline.At other timesit hasits body fluffed out and the tail down and closed,in the posturenormal for self-preening.It often tilts the head more and more at each peck of the groomer,freezing in each new positionlike a man in a barber'schair. Oddly, the groomedbird freezes its headin positionand ruffs the crown whetherthe mate is peckingits head or someother part of the body. Once a femalesidestepped, reversed, and ruffedher headas the male flew to a positionbeside her. He looked about, chirpedrepeatedly, and hoppedto preen only a centimeterfrom her. She looked down, her head fluffing di- minishingslowly. When he fluttered 0.5 meter away, she twice scratched her head over the wing, did a right sidestretch, and startedpreening again as if suddenlyreleased from forcedimmobility. The bird doing the groomingextends the joints of the legs and stretches the neck and head (extendsthe head on the neck, extendsthe joints of the neck,and lowersthe neckon the body). Often the headangles upward from the neck rather strongly. Since the body of the groomeris angledupward, the femora are probably extended. The legs of the groomer are splayed, anotherindication that the femora are extended. The other joints of the legs oftenseem to be extended,but I am uncertainwhether the feet and phalanges are normallyflexed or extended.The groomerruffs the crownup and down, but the body staysrather sleekedand the back patch seldomshows. If the groomeris also sitting and loafing, the body is fluffed as is normal during resting.The tail is closedand downas in the standardposture. At times one bird grooms the other then is groomed in return. The invitation-to-groomposture sometimes comes from the approachingbird rather than from the sittingbird. Groomingdoes not alwaysfollow the invitation posture. Once both birds went into the head-ruffed, neck-archedgrooming postureand froze, so that neither groomedthe other. Once a female ruffed her head and hoppedto a positionbeside her mate, but sheflew off when he continued to preen. As is sometimesthe casewithout grooming,the two may sit and preen (autopreen) next to each other after or before grooming. At times the two were so closethat one of the pair hit the other with the tail when it reversed 44 ORNITHOLOGICAL MONOGRAPHS NO. 10 directions.On one occasionthe female did a full right stretchand bumped themale off his perch with her wihg. More often, however, one bird hops a few centimetersaway from the other beforepreening, and preeningbirds are fartherapart than groomingbirds. Somefield notesfollow. In one typical sequence,the pair ignored the peepingof their nearly grown juvenile as the adult female twice flew up to the distanceof one body width or lessfrom the male. The two stood together each time, preening a little. The secondtime the female turned and nibbled at the head of the male. She then jabbed strongly, but he just ruffed his head and pointed his bill down, freezing with head somewhat tilted away from her and lowered. When she, in her lower and half-sittingposition, stopped nibbling his lower face he returned to a normal pose,then nibbled her head a few times. She startedpreening the wing toward him, brushinghim almost every time she extended it. He started to preen, but after taking a few blows of her wing he flutter-hopped over her to her right side. They stood facing different directions instead of the same way after he hopped back closer, to one body width from her. Both resumed preening; but the male had gained little, becausehis mate now preened her right wing, hitting him with it as often as she had been hitting him with her left wing before. At timesthe male or femalehops over the back of the other, which ducks the body and ruffs the head. Both chirp frequentlyduring such leaps and during grooming. Occasionallya bird groomsa young bird, so the behaviorpattern is not limited to courtship. Once I saw mutual groomingbetween mates after I flusheda male off the nest, but in generalthere is no groomingwhen birds haveeggs or nestlingsto carefor.

PAIR FORMATION Pairs were alwaysformed when one memberwas on territory. No wanderingbird held a mate, althoughwandering males chirp and feed femalesat timesand occasionallytravel with them for a few days. A territorialmale that haslost his mate or is waitingfor his first mate sings persistentlyand loudly,then flirts with and feedsany trespassingfemale. Pair formationis accomplishedby a few daysof repeatedcourtship feeding, flirting, and association of the two birds. Normally the new mate is a wanderingyoung female, six to twelvemonths old, or a wanderingold female that has lost her mate and desertedhis territory. In severalcases, females under six monthsold chirpedand were fed by unmatedmales actively seeking mates, but did not form permanentpair bonds. Oncea male seekinga mate pushedsuch a youngfemale off her perch with his breastas he chirpedrapidly, but she scarcelyreacted except to eat his gift. In general,females under six monthsof age act like juvenileswhen courted. A female that losesher mate generallywanders off the territory in a few daysunless she pairs quickly with a wanderingmale. One suchfemale flirted 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 45 with a wanderingmale by serpentine-singingand flying near him when he captured an insect. Mated birds sometimescourt birdsof the oppositesex to whichthey are not mated. Males whosemates are on the nest generallychallenge and chase trespassingfemales the first few minutes,but chirp and foragenear them if they evadeattacks and persistentlychirp nearby. Once a trespassingmated femalemanaged to steala spiderfrom a serpentine-singingneighboring mated male. In anothercase, two malesexchanged mates. A male that has losthis mate occasionallysteals the mate of one of his neighbors.In all such cases, the male presumablycourts an alreadymated bird. There was no caseof polygynyor of a wanderingfemale displacingan adult mated female, presumablybecause the mated female normally drives any new female away or becausea male with a mate is usuallyantagonistic. One immaturefemale repeatedlytried to take food from a mated male, but he jumpedinto more and more uprightand challengingposes each time she grabbed;he did not attack her, but flew over and fed his own mate. Another wanderingfemale chirped at a challengingmale and kept comingto him, but the residentfemale challenged and startedthe residentmale's challenges anew; bothdrove off the wanderingfemale. A matedfemale may toleratea trespassingmale rather readily,as long as neither her mate nor his mate is present. In fact, a trespassingmale often seemsdominant if the residentfemale is alone. Shekeeps out of his way, and the two chirp back and forth as do pairing birds. Even wanderingmales, without mates or territories, sometimestake the best sites at swarms of ants and restrictresident females to poorersites. One preeningresident male went into an uprightpose but did not call when a wanderingmale came up and displacedthe residentfemale. The residentmale than returnedto preening, leaving his mate to shift for herself as the wandering male took the best foraging sites. A wanderingmale often uses chirping or serpentine-singingwhen he movesin near a residentor wanderingfemale. Wanderingfemales often chirp as they move in near residentor wanderingmales. At times the wandering malesor femalesare drivenaway, but oftenthe chirpingcauses the aggressor to compressthe body feathersand return to foragingor even start chirping. After a few daysof pair formation, the paired birds wander togetherunless incubationor other duties interrupt. At times the two forage 10 to 30 m apart. If disturbedby a human or a hawk, a memberof a pair is likely to chip or sing until the mate answerswith similar notesor faint-songs;then the two move together. The two may loud-singto each other as they wander separatelyin the morningas if lookingfor a swarmof ants,and in the evening as they wanderbefore bathing or roosting. If they are closetogether, they chirp or faint-singnow and then. If one bird finds a swarm,it stopssinging 46 ORNITHOLOGICAL MONOGRAPHS NO. 10 and the other circlesin to it, singinguntil there is an answerfrom the foragingbird. Both chirp as the mate movesin and foragesnearby. One femalefluttered the tips of her wingsbriefly, like a youngbird, as sheflew to the limb besideher mate. At timesthe male startsrapid chirpingor amorous serpentine-singingas the female approaches,even thoughhe doesnot have food for her. At times there are faint snarlsamong the chirpingnotes when the pair work lessthan a meter apart,but supplantingsare rare. One male chipped and snarledat a new matefor a minuteafter shehad panickedand chipped; when she chirped,he graduallysubsided, looked about, and scratchedhis headover his wing;later he fed her. The male of a pair with youngin the nestchirped faintly and frequentlyand utteredbeeee'tree serpentine-songs as both foragedbusily; he supplantedher twicein salliesto the ground. She ruffedher head and froze on thefirst supplanting.Later theyperched a body width apart and chirpedbusily, both in ruffed-headedand slow-movingposes asin groomingbehavior. One othermale supplanted his mate when she did a sallyingcatch four m abovethe ground. She froze, flittingher wingsand flickingher tail. Finally she startedtentative chirps and he chirpedback. Periodicsnarls and supplantingsare more commonif the two have other mates. One femaleflitted her wingsand snarledfaintly as her mate worked 0.6 m from her. Flitting the wingsand flicking the tail are often reactionsto a matewithin 0.3 m, althoughin Hediger'sterms ( 1950:111 ) SpottedAntbirds are almost"contact" birds comparedto antagonistic"distance" birds like BicoloredAntbirds. The latter supplantmates closer than a meter away, and preenmore than a bodywidth apart,but SpottedAntbirds tolerate the matesso readilyit is difficult to tell whethermales or femalesare dominant. In the few caseswhere a SpottedAntbird did retreat as its mate came up, the male was alwaysthe displaceror supplanter.Normally the femaleand malewander more than a meterapart when foraging, so that the two seldom go for the sameprey item or competefor the sameperch.

COPULATION Courtshipfeeding always precedes copulation in SpottedAntbirds. After a numberof courtshipfeedings in the courseof a few days,the femaleno longerflees with the food as soonas shegrabs it from the male'sbeak. As shechews it with head down besidethe male, he tries a quick hop to her back. He is chirpingrapidly, as he has been sincehe gave her the food. At first shehops away or leaveswhenever he triesto hop on her. His chirpscome lessand lessrapidly. He looks at her, then returnsto foraging. Later she squats,with legssplayed and the feathersof the belly lowered,even as she takesthe food (Figure 10, a). Finally shestays, and copulationensues. 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 47

Figure 12. Copulation by Spotted Antbirds, from field sketches. a, Male flutters wings as female holds food that he gave her. b, Male flutters off as female rises,food still in her beak.

The female croucheson the limb and holds her head low, the food still in her beak. The male balanceson her back, sittingon his somewhatdiverging tarsi and on his under-tailcoverts (Figure 12, a). At first he shiftsthe feet back and forth. His feet are at the sidesof her back, just behind and above the insertionsof her wings, apparentlygrasping her back feathers. He is in a somewhatupright position, his headand bodyrather strongly fluffed. He has the neck very stronglyretracted, the head archedon the neck, and the neck lifted on the body. Thesethree movements, at lesserintensity, are characteristic of nearly all the courtshipencounters in this species.As his tail goesto one side of hers, he tilts backward so that the cloacae come into contact. His ventral apteriumis open, as he is fluffed very strongly,and apparentlyis inflated also. He usuallyflutters the wingsrapidly and regularly,perhaps for balance. The fluttering is somewhatpeculiar, suggestingthat certain jointsof the wingsare extendedor flexedmore than in normalflight. If the wings are flexed in the sameway as the legs, the carpal and perhapsulnar sectionsshould be extended and the hurneral section flexed, but this is uncertain. Most of the displayof the wingsis in the distal sections,however. As the two copulate,the chirpsof the male deepenand slowto a pip-pip-pip seriesat three or four notes per second. After copulation,the fluffed male (Figure 12, b) sometimesflutters off to a perchnearby. His flightis still fluttery,suggesting some flexion of sections of the wing as a holdover from the copulatoryactivities. The female, still crouched,finishes chewing and gulpingthe prey. At other times the female slipsout from under the male, sometimesso rapidly copulationwas probably not complete. In one such casethe female did not rotate her tail so his tail couldgo past hers,though he sat on her back and flutteredhis wingsfor 10 secondsas he movedhis tail oneway andthen the other. In anotherincomplete 48 ORNITHOLOGICAL MONOGRAPHS NO. 10 copulation,the male did not flutter his wings; after the female slipped out from under him he wiped his beak. The male often perches,fluffed and head retracted,on the perch the femalewas on. Sheworks intentlywith the food and ignoreshim. Preeningsessions sometimes follow copulation.The male is more likely to resumeforaging and the femaleto preen.

THœ NEST SITE AND NEST Spotted Antbirds serpentine-singand wander, flicking the spread tail frequently,when searching for nestsites. At timesone alightson the edge of an old nest and looks down into it. One male fed his mate and then went to an old nest and sat in it severaltimes, turning one way and then the other. Another male, apparentlyunmated, also visited an old nest and flicked his tail as he looked down into it; he gave serpentine-songs,then chirped, but soon rejoinedthe bird flock he had been with. One female visited an old Slaty Antshrikenest and peckedat loosestrands, perhaps searching for nest material. Skutch (1946: 18-19) has noted that other antbirds are interested in old nests,and that their visits to nestsof other speciesmay lead to ornithologists'mistaking ownership of antbird nests. One pair with a successfulnest (young left 10 August 1969) had small youngin the samenest 13 Octoberbut lost them by 21 October,then laid eggsthere April 29-May 1, 1970, but lost young between20 and 27 May (observationscourtesy of Robin Foster). One pair nestedin exactlythe same site as had a completelydifferent pair the precedingyear. Thus, old nestsor nest sitesmay be usedat times. Usuallythe bulky,leaf-decorated nest cup (Figure 13) is suspendedbetween two small divergingor parallel twigs of a slender sapling. Although the verticalsapling is normallynot stronglyincorporated into the sideof the nest, the nest is only rarely placed in a fork or betweentwigs more than a few centimetersout from somevertical support. Perhapshorizontal twigs farther out bend too much with the weight of the nest, especiallythe thin and weak twigsone usuallyfinds low in the forest. Nests are usually in fairly sparseundergrowth, with clear views in every direction. However, at times the nest is near patches of shrubs or near a sharp slope of a gully or stream, so that visibility is blocked on one or two sidesor the nestis hard to see againstthe ground. The nest of the SpottedAntbird is, next to the nest of the commonSlaty Antshrike,one of thosemost frequentlyseen in the undergrowthon Barro Colorado. The former nest is a thick and bulky cup of rhizomorphs("fungal horsehair") with pendentor plastereddead leaves. It is rather irregular outside,and looksmore like a pile of debristhan doesthe thin and neat cup of the antshrike.Perhaps the pendantsand plasteredleaves help concealthe nest of the Spotted Antbird somewhat,but it is rather conspicuousto a 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 49

;i '-. 7..

.•... ..•...... • •, /•, •- • • .. t•:'," -.... ', • '- •. ' ":' .• ......

•. - -. .• .-' • .-•'" . •.• .'• :: -. Figure 13. Nest of Spotted Antbird. human. Internally, it is a neat cup fined with black rhizomorphs. Much of the body of the nestis made of black rhizomorphsand other slenderstrands. Three nests that Skutch (1946: 19) found on Barro Colorado were rather similar, although one nest was composedmainly of slender pistillate inflo- rescences of Myriocarpa izabalensis (Urticaceae) and included few rhizomorphs. He recordsone nest as having an inner diameter of about 6.3 cm and a depth of 7.0 cm. The 120 nestsI had locatedby I September1970 were 0.3 to 1.7 m above the ground;the medianheight (for 113 nests,counted as 115 becauseone was usedthree times) was 0.7 m, or somewhatabove the usualforaging height. The mean height for these 115 nestingswas 0.78 m; the mode was 0.6 m (20 nestings). Four of the five nestsabove 1.4 m were on steepground or above gullies.

BUILDING THE NEST

Both male and female build the nest. The male did more work than did the female at the nine nestsat which I watchedconstruction, by 48 visits to 16 (3 to 1 ). At severalnests only the male visitedduring the brief periodsI watched. At one nest where I watched construction from a blind on 10 July 1966, the ratio during 113 minutesof watchingwas 21 visits by the male to 10 by the female. Skutch(1946:19) recorded19 visitsby the male and 13 by the female during two hours he watchedfrom a blind. He noted that the male seemedmore activein shapingthe nest. The female generally is more disturbed than the male when one watches without a blind. But both birds chip, flick their tails and flit the wings,and hesitatebefore going to the nestif the observeris in plain sight. Somebirds startedfaint snarlsas they flitted and flicked near me and the nest. From behind a tree at the above- 50 ORNITHOLOGICAL MONOGRAPHS NO. I0 mentionednest on 9 July 1966 1 watchedboth the nestand the birdsattending a passingant swarm and the female made only 4 visits to the male's 11. Probablythe ratio of visitsby the male to thoseby the more timid female is more likely to be 2 to 1 or lessrather than 3 to 1 when the birds are not disturbed. The one pair watchedfrom a blind generallyworked without alarm notes, but the male often approachedwith a seriesof chippingnotes. Alarm calls of a distantBlack-striped Woodcreeper and Plain-brownWoodcreeper caused the femaleto crouchon the nestand look up nervously.Later the male chipped as he approachedand left and was excited becauseof the continuingalarm calls of a Black-stripedWoodcreeper. However, as the woodcreepercon- tinued calling the pair worked even more quietly than before. After the woodcreeperhad stopped,the female SpottedAntbird gave a few chipping notesas sheworked. Other birds,watched without a blind, sometimeschipped as they workedon the nests. A female Chestnut-backedAntbird, "pounding"her tail (see p. 10), flew to a point below the nest and staredup at the male SpottedAntbird as he workedthere. She hoppedup the saplingand peckedat the male; he spread out hisbody and wings and fluttered the wingsas he peckedback. Shedropped to the ground as her mate came up; both poundedtheir tails as the male SpottedAntbird froze on the nest. The female then hoppedup again and drovethe male SpottedAntbird off the nestin a flutteringfight. Four minutes later the pair of Chestnut-backedAntbirds, now 5-7 m from the nest and foraging,ignored the returningmale SpottedAntbird. As is discussedlater, antagonismof Chestnut-backedto SpottedAntbirds both at and away from swarmsis probablyrelated to foragingcompetition, as the Chestnut-backed Antbird usesdifferent nest sites (Skutch, 1969). Usually birds buildinga nest tug and flutter as they pull off rhizomorphs and other thin strandsfrom trunks,limbs, and rotting logs 1-50 m from the nest. At times skeletonizeddead leaves,some so decayedthat only the midribs are left, are pulled from piles of trash near or on the ground. The birds seldomvisit the ground or move above 2 m from it. Ordinarily few leavesor strands,sometimes only one or two, are carried at a time. Once, however,a male carried a large tuft of rhizomorphsand left part of it on a nearbypile of debriswhile he took part to the nest. Then he returned,picked up the rest of the rhizomorphs,and took themto the nest. At the nestI watchedfrom a blind (08:30-09:46, 16 July 1966), the 16 intervals between 17 arrivals of the male with material varied from 2 to 9 minutesand averaged4.7 minutes;the 9 intervalsbetween 10 arrivalsfor the female varied from 1 to 13 minutesand averaged6.0 minutes. The 26 total intervals between visits of male and female varied from 30 seconds to 9 minutesand averaged2.8 minutes. 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 51

Often one memberof this pair arrived with material as the other worked on the nest. At suchtimes, as when they wanderednear each other looking for material,the male serpentine-sangor chirpedfaintly; the female often chirped. At times the male serpentine-sangor faint-sangto the female's faint-songsas he worked on the nest and she hunted for material 1-30 m away. When one bird first arrived as the other sat on the nest, he or she chirpeda few times or silentlywaited a few secondsuntil the bird on the nestleft. At timesthe arrivingbird hoppedup the stemof the saplingor waitednearby, but generallyit alightedon oneor the otherof the two horizontal twigssupporting the nest. In thesecases the arrivingbird alwaysalighted distallyto the bird on the nest, generallywithin a centimeteror two of it, as it quickly finishedwork. Once, eight minutes after the attack by the femaleChestnut-backed Antbird, the male SpottedAntbird gavefaint snarls from the nestas his matecame up; the pair of Chestnut-backedAntbirds were stillnearby. Otherwise there was no signof antagonismbetween the members of the pair of SpottedAntbirds. Usuallythe arrivingbird depositedits strandson one or acrossboth of the twigssupporting the nestas soonas it alightedor as soonas it hoppedin to the nestwhen its matedeparted. Then the bird lookedabout briefly, pecked at or tuggedat the strandjust deposited,pulled it around the stem or one or both twigs, and departed. Pulling and pushingstrands with the bill from aboveseemed the main method of constructionat this stage. At times the bird depositedmaterial and left within a few seconds.Later in the hour I watched,as a thin cup or cuppedplatform of rhizomorphsand a few skele- tonizedleaves began to take shape,the female sat directly in the nest and pressedher body downinto it. Sheapparently kicked at the sametime, thus enlargingthe insideof the cup and tanglingthe strands.Earlier in the morning the male arrangingmaterial had squattedon the southtwig, but not on the nest itself, which was then a slightbridge or saddlebetween the twigs. As he half-sat and pulled material one way and then the other with his bill, he flicked his closedtail as if excited. He first steppedinto the nest and did sit-kickingin the nestseveral times at 08: 39. The morningbefore, this nest had consistedof just a few strandson eachof the separatetwigs supporting the nest.

THE EGGS AND INCUBATION Thisnest received the first egg15 Julyand the second17 July. In thisand severalother nests little newmaterial was added during the two or threedays before the first egg was laid. However, I flushed the male off this nest at 07:45 on 13 July and at 08:18 on 14 July,when a new strandwas overthe nestcup, so thereis someaddition even on the daysbefore the eggis laid. 52 ORNITHOLOGICAL MONOGRAPHS NO. I0

In onenest, barely started on 30 June,the first eggwas in the neston 6 July andthe secondon 8 July. At threeother nests, two daysintervened between the laying of the first and secondeggs. Probablynest constructionand the followingless activeperiod generallytake about a week, and eggsare laid twodays apart. The BicoloredAntbird and Slaty Antshrike also lay eggstwo daysapart, suggesting that this may be regularamong antbirds. In two casesfor which data are available,female SpottedAntbirds visited nestsand laid eggsbetween 07:15 and 08:18 and between07: 30 and 09: 00. The two eggsare whitish,streaked and splotchedvery heavilywith dusky brown to chestnut.I had observed84 suchclutches by 1971; two eggsin one other clutch were stained chestnut all over. Four other nests I found, on whichpairs were incubating,held only one eggeach. In mostcases there were two youngin nestswhich containedyoung birds, but in severalcases therewas only one youngbird. In caseswhere only one egg or youngwas presentthe originalclutch may have been two, and one egg or younglost. Skutch (1946: 19) reported two clutchesof two eggs each from Barro Colorado, motfled with umber on whitish; a set he measuredwere 23.0 x 15.1 and 23.0 x 15.9 mm in dimensions.Four eggs I measuredwere slightly smaller,averaging 20.9 x 15.0 min. Thesefour eggsweighed less (average, 2.15 g) than did 6 eggs(average, 2.6 g) not so measured. Both sexesincubate, as is usual in the Formicariidae (Skutch, 1969: 182). Both males and femaleshave vascularized,bare brood patches. Regular incubationbegins when the secondegg is laid. I havenot seenbirds on nests whenonly the first eggis present,but theremay be occasionalincubation. In one caseyoung hatched a day apart, suggestingthat the first egg may have beenincubated alone for sometime; however, the lastegg at thisnest hatched after 16 days,a rather long incubationperiod. Often one egg in a nestwill be wet from rain. At times,however, even two eggsmay be wet from rain if bothbirds happen to be off the nestwhen rains hit; apparentlythey do not attempt to return rapidly. Judgingfrom birdsobserved at ant swarmsand flushedoff nests,each bird incubates one to four hours at a stretch. The male incubates for an hour or two in the earlymorning, after the femalehas incubated all night. Only the femaleis at a swarmof ants beforeabout 08:50 to 09:30 if the pair is incubating.The female then stayson the nest an hour or two. The male oftenincubates the rest of the morning. In the afternoonthe birdsalternate. The femaleusually has three or four sessionsduring the day, totallingabout four to six hours (midmorning,midday, midafternoon, late afternoon) and themale four or five,totalling about six to eighthours (early and late morning, early and middle and late afternoon). Skutch (1946: 20) found a similar patternat a nesthe watchedfrom one noonto the next,but only two sessions by themale and three by thefemale; the male was on 338 minutes(47.0 per- 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 53 cent), the female on for 327 (45.5 percent), and neither on for 54 (7.5 percent). The femalewas on at night. At that nest,the male arrived with excited chippingnotes for bothincubation periods and the female gave sharp chipping notesafter she settledon the nest one time; otherwisethey were silent. The maleleft the nest,chipping, for eightminutes during one session. I watchedfrom a blind at one nest after flushingthe male off at 14:24. The femalecame up with loud chippingat 14:54 and continuedloud calls evenas sheflew to the southtwig supportingthe nest. Shelooked in, flicking her closedtail, and hoppedon at 14:55. She continuedloud calls as she settled,then was quiet. Except for ruffling her head and looking about, turningher headthrough a smallangle each time, shewas quiet until the chipping notesof the passingmale were heardat 15:44. Then shelooked about actively,hopped up and over the edgeof the nest,and flutteredaway low and silentlynortheast toward the male. Incubation,from the layingof the last eggto the hatchingof the second young,required 15 daysat thisnest and at one other,but 16 daysat a third nest. Skutch(1946: 20) predicteda 15-day incubationperiod for Spotted Antbirds,based on hisknowledge of the incubationperiods of otherantbirds.

I watchedfrom 07:40 to 15:41 on the day of hatching(1 August) at the above nest. The male, on the nest when I arrived, stayedlow for 10 minutes after I arrived. At 08:18 and 08:34, after much looking around, he rose and poked down in the nest. From 09:02 to 09:05 and 09:14 to 09:15 he rose and looked beneath himself several times, champing his bill and ruffling his head each time. At 09:15 he rose again, grabbedhalf of an eggshell,and flutteredoff eastwardwith it. One tiny blackishyoung was in the other half of the eggshell.At 09:20 the male came back, calling peep-peep- peep! repeatedly; he flew to the nest and half-sat, flicking his closed tail as he pecked down into the nest. At 09:22 he flew off silently;at 09:26 he returnedsilently to a vertical stemnearby, then flew to the twig by the nestand starteda chirpingseries. He stoppedand froze as a coatimundi passed 15 m to the northeast. Finally he looked down and peckedin the nesta few times, twitchedthe wings apart a few times, fluffed out his underpartsgradually, looked about, looked down, and at 09:30 hoppeddown on the nest. When the female arrived at 09:36, they chirped faintly and he flew off past her. She camesilently to the edgeof the nest,looked down with tail closedand down, pecked into the nest, and a minute later flew off with the other half of the eggshell. Back she came to the nest edge, looked and pecked down a bit, then fluffed out her underpartsand shifted her wings back and forth and hopped down on the nest. She took a long time to settle, and her head was rather fluffed. After looking about very actively and giving faint snarlstwice, perhaps at my movementsin the blind, she suddenly(09:57) hopped out of the nest and grabbed a sowbug from the ground below. She seemed to eat it, but on her return she pecked down into the nest several times. She fluffed the belly feathers,flitted the wings a few times from her hunched position, and hopped on. She looked about more actively than she had when incubatingeggs on 27 July. At 10:18 she rose and poked under her, then settled. At 10:26 she rose as if pushed; 10:36, she rose, looked under her, picked up half of the secondeggshell, and flew low off northeast with it; 10:40, the returning male, calling peep-peep!loudly, took food to the nest; 10:45, the female, chirping, replaced him at the nest edge and fed the young. She flew 54 ORNITHOLOGICAL MONOGRAPHS NO. l0 off to a point near him. He salliedto the ground,took the youngmore food at 10:48, and hoppedon the nestto brood until the female arrived at 10:55 with still more food. At this nest,therefore, the two younghatched about an hour apart, and bothparents carried the eggshellhalves away. The youngwere fed lessthan two hoursafter hatching.The detailednotes continue. From 10:43 to 15:41 there were 17 feedings,8 by the female and 9 by the male. The 16 intervalsbetween feedings varied from 2 to 76 minutesand averaged18.2 minutes; eight intervalsfor the male varied from 5 to 83 minutesand averaged36.5 minutes; sevenintervals for the female varied from 10 to 110 minutes and averaged 40.3 minutes. The male alsobrought food at 12:02, but droppedit below the nest;when tossingleaves therebrought nothing, he left and returnedat 12:11 with anotherprey item. Three times the male left the nestafter feedings,at 10:43 and 12:47 and 13:28; the female left the nestafter the 10:45 feedingonly. At othervisits the feedingbird fluffedout and hopped down to brood the young. The male brooded five times, 3 to 41 minutes at a time (average, 21.6 minutes) and was on the nest when I left; he brooded or was feeding.young 116 minutes and was away 174 minutesbetween 10:45 and 15:35. His eight periodsaway from the nest varied from 2 to 76 minutes (the latter was the time he returnedafter 67 minutesbut droppedthe prey). The female'sseven periods away from the nestvaried from 1 to 44 minutes. She brooded seven times, 1 to 66 minutes at a time (mean, 20 minutes); she brooded or was feedingyoung 148 minutesand was away 142 minutes. One parent or the other was thus at the nest 264 out of 290 minutes, or 91 percent of the time, between 10:45 and 15:35 on the day of hatching. Skutch (1946), as has been noted, found that the two parents incubate 92 percent of the time. Except for two times (12:47 and 13:28) when the male left directlyafter a feedingand the one time (10:45) when the female did so, the feedingparent stayed and broodeduntil the other parentarrived. Several times the male gave a seriesof chippingnotes as he approachedthe nest, but at other times he arrived quietly or gave faint chirps or faint-songsto the female. He sometimeschipped near the nest, but was silent on it or gave faint chirps as he bent down to the young. She gave faint chirpsor faint-songsto him when he arrived as she brooded,or she departedsilently past him. She chirped or faint-sangor was quiet as she arrived. He was silent or chirped faintly as he left past her. Neither was noisy on the nest bush,although the male twice continuedchipping until he was a few centi- meters from the nest bush. Once the female gave a faint snarl at a Slaw Antshrike foragingpast. Usually the parentsarrived low, saplingto saplingnear the ground, and circled somewhatbefore they came in to the nest bush and then flew or hoppedup to the nest. The prey fed the young was mostly very small, less than 5 mm long; some items seemedto be sowbugs.The parent half-sat and lowered its closedtail as it hunched over the nest from the twig supportingits south edge. At times a few chirpspreceded the feeding,which was alwaysrapid. The parent then champedthe beak and looked about. At times it poked or peckeddown at the young, but I saw no sign of eating or otherwisedisposing of fecal sacs. A minute or two after arriving, it either flew off silentlyor spreadthe ventral feathersgradually, rearranged the wings,and hoppeddown to settle over the young. Feedingand broodingbehavior of adultsat two othernests on the day of hatchingwere generallysimilar. The adultsgenerally brooded young after widelyspaced feedings. Other detailednotes follow. 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 55

At one nest I watched from 16:00 to 17:20 (24 July 1961) and took flash pictures of a pair feeding young in Day 4 (i.e., hatched somewhat over four days earlier). This pair approachedand fed the young silently, and did not stay to brood. The male gave a few chirps as he departed past the incoming female on one visit. They approachedthe nest by short flutters, flew up to it from below, and dropped near the ground on leaving. Seven intervals between eight feedingsby the female varied from 1 to 39 minutes (average, 9.3); she fed the young five times between 17:03 and 17:13 and was foraging very close to the nest. Four intervals between five feedingsby the male varied from 1 to 21 minutes (average, 10). Twelve intervals for both parents ranged from 1 to 24 minutes (average, 5.5). Once the prey was a 2-cm caterpillar; the female had to rechew it after placing it in a young gape. The female carried away a fecal sac after this feeding, at 16:13; the male carried away fecal sacs after feedings at 16:37 and 16:50. Feedingsand broodingsat five of six other nestswatched by myself and Yoshika Oniki in 1969 and at one nest watchedby StephenKistler in 1970 were similar. (At the sixth nestin 1969 the pair were very disturbed,either by my blindor by snakes[see below], and feedingswere irregular;9 intervals betweenvisits of both parentswhen the two young were in Days 2 and 3 averaged34.5 minutes.Then one young disappeared, and 5 intervalsbetween visitsto the singleyoung on Days 4 and 5 [beforeit disappeared]averaged 18.7 minutes.) At the six normal nests,feeding rates increasedgreatly after Day 0, but decreasedsomewhat on the last day in the nest: the averagefor 19 intervalsbetween feedings of two youngby bothparents was 18.0 minutes on Day 0, the day of hatching;43 suchintervals averaged 11.0 minuteson Days 4 and 5; 67 suchintervals averaged 12.5 minuteson Days 6 and 7; 88 intervalsaveraged 5.6 minuteson Days 8 and 9; and 38 intervalsaveraged 9.4 minuteson Days 10 and 11. High feedingrates on Days 8 and 9, and occasionallyon Day 4 (seeabove), were alwaysassociated with parents'foraging near the nest. Parentsnormally stayednear the nest on Days8 or 9 to Day 11, andoccasionally did soon other days. They did not alwaysfeed youngrapidly when stayingnear the nest, however. We recorded185 feedingsby males and 108 by femalesfrom Day 0 to Day 11. Femalesfeed the young as assiduouslyas males (44 by males/47 females)until youngare in Day 5, but thenbecome less active (141 males/61 females) from Day 6 to Day 11. However, one female fed as much as the male on Day 9, and the femalealways resumes full activityas soonas the first youngleaves and shehas to feed either the one in the nest or the one outside of it by herself. There was little daytimebrooding at the six normal nests,except once during a period of rainy weather. The last broodingrecorded was for three minuteswhen young in one nest were at Day 8. There was fairly regular daytimebrooding at the disturbednest in 1969 whenyoung were in Day 2, part-timebrooding on Days 3 and 4, regularbrooding on Day 5, and none 56 ORNITHOLOGICAL MONOGRAPHS NO !0

Figure 14. Female Spotted Antbird waiting for possible fecal sac after feeding nestlings. This nest was used three times in 1969-1970, only the first time successfully. on Day 6. Flushingparents off other nestsalso suggeststhat broodingde- creasesrapidly after the first few days, concurrentlywith the above-noted increasein feedingrates (primarily becauseof an increasein feedingby the male). Females regularly brood at night; one arrived at 18:08 when young were in Day 6. During the first hoursafter the younghatched, no fecal sacswere removed (none in 22 visits at three nests). On Days I and 2, fecal sacswere eaten before the parent brooded young on three of eight visits. At one nest, only one fecal sac was carried away in 1 l feedingson Day l l, just before one young left. On Days 3 to I I at other nests,however, fecal sacswere carried away by a parent, its neck unusuallyextended, after 38 of 137 feedings,or 27.8 percent. One female carried a sac to a perch 13 m away and shook her head after dropping the sac. Often the parent waits a few secondsafter a feeding,champing the beak and peckingdown into the nest, if no fecal sac appears(Figure 14). For other views of attentivebehavior see Plate 2.

I watched at one nest between 08:43-13:45 on 22 August and 05:45-07:13 on 23 August for the departure of two young. The male started chirring and chipping near my blind at 08:58, attracting the female to give faint snarls. Then they chirped until 09:12, which started loud peeee! notes from one young in the nest. At 09:15 one parent moved to the palmettos under the nest, the young fluttered and hopped up on the nest edge, and immediately fluttered to the ground and away along it. Its peeee! notes alternated with the serpentine-songsof the male that was leading it away. The female PLATE2. Spotted Antbirds at the nest. Above, male looks at young after a feeding. Below, female starts to carry fecal sac away. 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 57 stayed and foraged within 30 m of the nest the rest of the observation period. She alone fed the young remainingin the nest, and the calls of the male came only from near a treefall some 50 m to the east. The young in the nest gave some faint-peepingnotes. The flutter of wings from the female'sarrival at a low perchnearby often startedit to utter a noisysqueaking. After feedingit sometimesgave faint-peeps and preeneda little. Once whenthe female stayed three minutes on the nest edge the young peeped and preened busily, then gaped and squeaked into her face. From 11:02-11:20 a Keel-billed Toucan foraging near the nest and later the treefall caused repeated chipping from both birds. There was no feeding between 11:02 and 11:27, the longest interval until feeding slowed to three times per hour about noon. The female at times chirred when I moved abruptly in the blind, which was fairly thin. On my passingat 17:50, she chirred rapidly except during the few secondsI passed the nest. The next morning I arrived in the dark. The brooding female slipped off the nest in the semi-darknessat 06:15, and fed the young every 2.9 minutes (10 intervals) from 06:31 to 07:01. Only oncedid shecarry a fecal sac away. The young preenedactively, gave some peeping, squeakedand gaped before rapid feedings,and at 06:47 tried several times to flutter up on the nest edge as the female approached. At 07:01 she stayed below the nest after feeding; the young tried several times to hop on the nest edge, succeeded,and at once leaped to the ground below her. A few minutes later it was 0. I m up on a low twig there; the female chirred rapidly as I caught and weighed it. Thus the youngleft separatelywithout any practiceat flutteringin the nest, and only one parent fed the young remainingin the nest. This also occurredat two other nests;at one nestthe male stayedwith the remaining youngand at the othernest the femaledid so. Exceptingthe above-mentionedrapid sequence,30 intervalsbetween feed- ingsof the one youngremaining in the nestby a singleparent averaged 10.7 minutes(with the rapid sequence,40 intervalsaveraged 9.5 minutes). This is barelylonger than the averageinterval for two parentsfeeding two young in the neston Days 10 and 11 (38 intervals,9.4 minutes). The youngthat remainsin the nest (and probablyalso the one outside the nest) is fed almosttwice as rapidlyas it wasa few minutesbefore the first youngleft. However, the two youngin the nest are alsofed at this fast rate on Days 8 and 9, two daysbefore they leave.

GROWTH OF THE YOUNG The youngare very small, black, and lack down at hatching,as Skutch (1946: 21) hasnoted. They weighup to 2.3 g, probablynormally somewhat lessthan the 1.8-2.4 g of an eggbefore hatching (Figure 15). By two days and a few hoursafter hatching(Day 2), the quills of the remigesare appearing;the longest(number 4 from the end) is about 1.0 mm long (Figure 15). The younggrasp the nestlining weakly and peep faintly when removed. Theirbellies are relatively large and bulky, and the major process going on is plainly assimilationof food. 58 ORNITHOLOGICAL MONOGRAPHS NO. 10

15 30

• IO 20

½,.9 • 9 4 Z • / / PRIMARY

/% -5 5

0 I 2 3 4 5 6 7 8 9 I0 II AGE IN DAYS

Figure 15. Measurements of the longest primary remex, and weights, for young Spotted Antbirds.

Black pinfeathersare evidenton the wing covertsby Day 4, somerusty tips on Day 5, and brushy tips by Day 8. The primariesare openingto brushytips on Day 8, and have vaned tips along20 percentof their lengths on Day 10. Long pinfeathersare on the dorsaltract by Day 4, brushytips on Day 8, and long vanedtips from then on. By Day 8 there are short tips on the ventral feathers. Thus, most of the opening of feathers occurs on Day 8 and thereafter. On Day 2 the eye slits are startingto appear. By Day 4 the eyes are held partly open, and on Day 7 and later the young usually keep the eyes open when held in the hand. Young gape and squeakto the touch of the hand until Day 5 or 6, but crouchin the nest and huddlein the hand thereafter. On Days 8 and 9 and thereafterthey struggleand attempt to jump out of the hand or weighingapparatus. They clutch the nest lining little when removeduntil Days 5 and 6; by Days 7 and 8 their feet are well developed and clutchthe nestlining strongly. Squeaking changes to peepingnotes when 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 59

Figure 16. Young Spotted Antbirds on Day 4; from the nest in Figure 14. removed on Days 4 and 5, then to gradually louder pee-pee-pee-peenest- leaving notes from Day 7 on. As notedabove, young squeak noisily as the parent approachesand feeds, peep nest-leavingnotes now and then, preen betweenfeedings, and--as in- dicated by approximately15 observations--hopout of the nest during or near the end of Day 11 (i.e., on the twelfth solar day in the nest or early in the morningof the thirteenthsolar day. just beforethe youngis 12 days old). One other young left late on Day 10, but the disappearanceof its nest-mate on or about that time suggeststhat a predator got one young and the other fled prematurely.

FLEDGLINGS

Soon after it leaves,a youngbird can flutter up to 5 or 10 m only a few times in succession,with some lossof elevation,then must hop-flutter rather 60 ORNITHOLOGICAL MONOGRAPHS NO. 10

Figure 17. Young Spotted Antbird, just out of the nest, on Day ll; from the nest in Figure 14. weakly over the groundor alonga stem. It can cling to a vertical perch soon after leaving, but tires rapidly. It often falls over the perch on alighting. When not being chased,it sits most of the time on a low, slendertwig near the ground, attended by one parent. Figures 17 and 18 show such young. The young often is in densesprouts, treefalls, and similar places where it is difficult to detect unlessone triangulateson its faint calls to the attending parents. A young just out of the nest or sitting on the nest edge is very poorly feathered. The tail is a set of brush-tippedfeathers about 1 mm long. Some- times the rear end of the body twitches,especially the white feathers of the undersides,as the wingtipsflit upward as if the fledgling is excited; in the hand one can see that it is flicking its tail. However, many fledglingsseem placid on a perch or in the hand, and struggleonly to stay on top of the hand or to escapeany dorsal touch during the first minutes after capture. Sometimesthe young bird gives a faint pee pee pee pee. The wings are rather well feathereddorsally, with bright buff or chestnuttips to three rows of wing coverts(the greaterprimary covertslack tipping) and on the leading edgesof the secondariesand the tips of the inner secondaries.There are no wing coverts under the wing. The remiges are all short, but the vanes are four-fifths or more expanded,so the shaftsare bare only at the bases. There 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 61

Figure 18. Young SpottedAntbird (c) and distractiondisplays by parents,from field sketches.a, FemMe crouches,snarling and showingwhite back-patch,near a fledgling. b, Male flutters along ground with tail spread and dragging,near fledgling. d, Female takes upright frozen posture after being flushed off nest and eggs. e, Same female in downward frozen posture at another perch a few secondslater. are enough unsheatheddull chestnut-browndorsal feathers to cover the back betweenthe wings when the fledglingsits hunchedup in its usual waiting posture. The crown of the small head, otherwiserather bare and blackish,is coveredwith a "crew-cut"patch of dusky,brush-tipped feathers. A few auricularfeathers have brush tips, but the faceis mostlybare. The bill is blackish,a bit soft and irregular and small; the not-very-prominentgape anglesare yellowish-white.The rather loose and downy whitishfeathers of the lower breastand alongthe sidesleave the ventral apteriumand undertail open and bare. The breastis well coveredwith brownish-smudgedsimilar feathers.The large bare pinkishshanks and the largefeet are slightlysmaller than adult size, but are well developedexcept that the slightly short toes grip weakly. I havenot found a way to distinguishthe sexesat this age. As in BicoloredAntbirds, one parentfeeds one youngand the other parent feedsthe other. If one youngis lost, its parentinfrequently helps the other parentfeed the remainingyoung. However, after a hawk killed one male his mate fed both youngbirds of the family. If there is only one young bird and the female is feeding it (6 cases), the male sometimesfeeds her and shetakes the food to the fledgling. If the male is feedingthe only young (18 cases), she flirts with him but is not fed until the youngis gettingfood for itself to someextent. The great preponderanceof male care of singleyoung perhapsarises from the greatercare of nestlingsby malesin the last few days, 62 ORNITHOLOGICAL MONOGRAPHS NO. 10 hence, the greater chancethat males will lead any singleyoung from nests or take them over later. If there are two young, one parent occasionally appearsand chirrsbriefly as the observerchases the youngof the chirring other parent. Generally the extra parent returns to its own distant young when the attendingparent stopscalling. At times the male and his young bird are 100 m or morefrom the femaleand her youngbird, especiallyduring the first few days,when the youngare not easilymoved. At timesthere is a permanentexchange of youngbetween parents; in one casea male took over an immature female and the female took over an immature male. However, amongfamilies with one youngfemale and one youngmale I notedno significant tendencyfor parents to feed young of the oppositesex. In 8 such families malesfed the young males while in 11 the femalesfed the young males. This wastrue eventhough older young of the samesex as the attending parent are more likely to elicit snarls,challenging, and supplantingsfrom the parent. As a fledglingwaits quietly, its own parent wandersabout foraging. Oc- casionallyit flies to the youngbird and feedsit. After capturingfood, the parent chewsit and startsserpentine-singing. The youngbird answerswith faint-peepingand startsflicking its tail andlooking about even if (or especially if?) it is on one'shand. If the youngdoes not answer,the songsof the wandering parent are louder and louder, and it stops chirping. Eventually the parentwanders past the callingor silentyoung bird and feedsit or leadsit off. Usually the chirpingpart of serpentine-singingpredominates as the parent flies up and feeds the young, althoughthere is not often a rapid seriesof chirpsas in feedingthe mate. The youngflutters and squeaksas it extends the neck and head just above the horizontal and gapes at the parent, which often recoils a little. Then the parent insertsthe prey into the open gape of the young. The parentwatches the gulpingyoung, then flies off. If the parent is performing"leading behavior," it staysa few metersfrom the young bird and chirps and serpentine-singswhile reversingback and forth and flicking the tail. Soon the youngis hoppingand flutteringover the ground or amonglow twigs after the retreatingparent. The latter feeds it whenthey have reached a newhiding place. At times the parent works a fairly distant swarm of ants and commutes 100 or 200 m to the youngbird, but generallythe parentworks near the young bird even if there is a swarm within 200 m. In somesuch cases the parents may have failed to discoverthe swarms,of course. As the youngbirds grow, they and their parentsare more likely to be at any swarmsof ants in the neighborhood.Within a few days after leavingthe nestthe youngbird gains the ability to flutter 10 to 15 m repeatedlyfrom one horizontaltwig to another without losing much altitude, even if it sometimeshas to flutter 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 63 and pull itself up on the percheswhen it alights. At this stagethe young can be led rather easilyby its parent.

DISTRACTION DISPLAYS AND PARENTAL ALARM Parentsflushed at nestsflee and chirr repeatedly.Usually the incubating bird fleeslong beforeone reachesthe nest,but at timesthe bird stayson the nestas one passeswithin a meter or two. At timesa flushedbird givesa brief snarl, but most quickly disappearsilently after the chirringsequence. A bird flushed off thc nest sometimesflutters low over the ground. One female,flushed off eggs,fluttered low andthen froze at 0.1 m abovethe ground for a minute (Figure 18, d); then she fluttered 10 m farther and took an upright, spread-wingedpose (Figure 18, e) on a vertical saplingfor a minute. The latter pose, but with wings spreadupward, was presented to somethingon the ground near a nest, where snakestwice attempted robbing(see below). In this casethe female startedgiving faint snarlson the nest, the male cameup and snarledfaintly, and she flutteredoff to give the displayand snarl faintly. Both birds lookedat the groundand snarled faintly. The nest was empty the next day. This pair often gave faint snarls on the nest; two days before the above incident the female gave thesenotes and then flushedoff as a snake (Pseustespoecilonota) reached up to rob the nest. (I collected the snake as it reached in.) Once I noted a Spotted Antbird awayfrom any nestgiving faint snarlsat a vine snake(Oxybelis sp.), whichis anothersnake that probablyrobs nests. Parent birds normally are quiet near their nests. Supplantingattacks on trespassingSpotted Antbirds often are silent,without the usualsnarling. Often the pair simplydisappears, leaving a good swarmof army ants to trespassers, if a human is nearby. If one approacheswithin 30 m of a fledglingSpotted Antbird or of nestlings 10 or 11 days old, the parent usuallycomes up and startschirring rapidly. There is usually tail-flicking, wing-flitting, reversing, and hopping from perch to perch. The throat may be puffed out. Betweenchirrs, the parent occasionallygives a faint snarlor two. At timesthe white centerof the back shows,another suggestion that the bird is partly in the challengingdisplay as well as in the mobbingdisplay. Chippingand panickingare rare, however. The parent wandersnear one in a quite distractingand conspicuousway. Often it movesto the side away from the youngbird, or cutsbetween one and the youngif one happensto go toward the young. However, the parent circlesabout so irregularlythat one seldomcan find where the youngbird is unlesshe backswell away and waits until the parent startssinging and the younganswers. Sometimes the parent startsleading the youngaway. If the fledglingis in the hand, it givesfaint-peeps and flicks the tail as soon as the parent startsgiving faint-songs, but it staysvery still if the parent is 64 ORNITHOLOGICAL MONOGRAPHS NO. 10 chirring. One faint-singingmale, which came with a caterpillarin his beak as soonas the fledglingstarted faint-peeping from my hand,was very agitated. He spreadhis tail and flickedit, spreadhis bodyuntil his flank featherswere over the wings,and flitted the wingsand pivotedback and forth as he gave snarlingand chirringnotes rapidly. The youngbird leap-flutteredhis way. When I chasedit, the male chirred and fluttered back and forth from one low perchto anotherbetween me and the young. He had his wingsdown and beatingstiffly in distractiondisplay at each perch. Once he flutteredalong the groundin a more typicaldistraction display, a "flutteringrun." This and severalother parentsfluttered past their hopping young in a "shielding flight" wheneverthe fledgingmoved. In one particularlybeautiful instance, a female knockedher flutteringfledgling from the air and hoppedcon- spicuouslyaway over the groundafter "continuing"its flight a meter beyond the point where the nonplussedand frozen fledglingtumbled down. The young stopsin such cases,and I generallyfound myself following the con- spicuouslyfluttering and callingparent. The "flutteringrun," noted also in other parents,involves spreading and tilting the tail one way and then the other as it is depressedto or nearlyto the ground. The white back patch is spread, and the wings, slowly beating and partly spreadto either side, emphasizeit as the bird sits on or flutters alongthe ground. The body feathersare spread,as are the throat feathers. At timesthe wingsbeat in rapid burstsas the parentwheels or croucheson the ground. At other times the parent performs the display from a low perch,or hopsup anddown onto low perchesas if runninghurdles as it flutters alongthe ground.One old malewho did an especiallygood series of fluttering runs alsodid several"shot-bird" displays, involving a slow forward pitch off a horizontalperch, clinging head down an instant,dropping to the ground 0.4 to 0.6 m below,flopping briefly, and then hopping away flapping the wings. Anothermale did the samedisplay, showing his white back as he did so. I havenoted this display also in Red-throatedAnt-Tanagers (Willis, 1961: 499), so it occursboth in songbirdsand suboscines.After distractiondisplays, as the parenthops around, the bendsof the wingsare held out, the crownfeathers go up and down,the tail is flicked,and the wingsflit. When young are a week or two out of the nest, they fly rather well and tend to fly higherand higher,even to 15 to 20 m up in vine-toppedtrees, if one chasesthem. The parentstays low in the undergrowthmost of the time, exceptto feed the young after one backs off. At suchtimes the parent flies around chirring,and does not go into distractiondisplays except for oc- casionalshielding flights. At timesI caughtboth the youngand the parent in a net when the parent tried a shieldingflight. Once a male flew at and knockedits youngoff a perchwhen it wasin plain view. 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 65

DEVELOPMENT OF YOUNG OUT OF THE NEST Young SpottedAntbirds grow rapidly and molt into adult plumageduring the first sevenweeks after leavingthe nest. (To get agesfor the following accounts,add 11 days if age since hatchingis desiredand 26 days if age sincestart of incubationis wanted.) During the first week, the tail grows until it is visiblebut only aboutone-third the adult length. The youngbird is still belowthe adultsize at the end of the week. The headseems especially smallor short-featheredalthough the dull throat and facial featherscover the bare areasby Day 7. Downy, brown-smudgedfeathers cover the breastand dingywhitish ones cover the lower underparts.The youngbird hidesand waitsmuch of the time, but readilyfollows the parentif it startsserpentine- singing.After the younghas been out of the nesttwo or three days,it is possibleto catchit only by chasingit severalhundred m, perchto perch, and shakingit off whenit startstaking high perches and freezing there. Duringthe secondweek out of the nestthe youngreaches nearly or fully the adultsize, to 17.7 g by Day 10 in one case,although the dull brownish headis still smalland ragged. The tail is up to three-fourthsthe adultlength by Day 14. The remigesare well developed,but the wingsare still bare of covertsunderneath. If the youngbird is a male, new, bright chestnutback feathersof the adultplumage are appearingunder the brownishfeathers of the juvenalplumage. New bluish-grayfeather brushes are also appearing amongthe brownishjuvenal feathers on the head. One can tell the juvenile malein the field by his verydark head. The femaleis alsogetting new head and backfeathers, but theseare approximatelythe samecolor as the juvenal feathersaround them. The chestsof both are still an unattractivesmudgy brown. The youngfly increasinglywell, so that one cannotcatch them by chasingthem. Theyflick theirtails and hop andflutter from perchto perch almostas well as do adults;no longerdo theytumble past perches or have greatdifficulty alighting on verticalor bendableperches. It is very difficult to find or seeyoung between Day 7 andDay 14, becausethey flee readilywith their parentsor hide very expertlyin densetangles. I have relativelymore observationsof olderand youngerjuveniles. During the third week out of the nest, the youngbird reachesfull adult sizeand tail length. At Day 2l the youngbird is in full post-juvenalmolt. The headis oftenvery raggedand in somebirds is juvenalbrown, but in most is approachingthe color of the adult. Males are gettingblue-gray heads and blackthroats, as in the adultexcept for dull brownfeathers on the crownsand auriculars.They haveone or a few new black-tippedwhite feathers breaking the unattractivesmudges across their chests.Their backsare mostlybright chestnut,as in the adults.Females are gettingnew buff-tipped whitish throat feathersand buff breastfeathers. Males and femalesare now easily dis- 66 ORNITHOLOGICAL MONOGRAPHS NO. I0 tinguished.Young of this age flee and hide rather readily, especiallyif their parentsare disturbedby the observer.One seldomsees them, exceptin mist nets. During the fourth week out of the nestthe youngare slowlygaining adult plumage. Young and adult femalesare not easily told apart, exceptby the indistinctsmudgy patching across the breastsof the youngfemales and by their pale gape angles. The smudgingon the breastof the youngmale is more conspicuousamong the new spotted black-on-white feathers. Head and backmolt are nearlycomplete by Day 28, althoughthere are someshort anddull juvenalfeathers breaking the regularoutline of the head. The young are flying rather well, and startingto forageclumsily on their own (as early as Day 24 in two cases). They play with smalltwigs, bits of leavesor whole dead leaves,peck at bumpson branches,etc. Their parentsstill feed them most of their food, but no longerdo they chirr persistentlyor lead them off from the observerso actively.It is thusrelatively easy to seeyoung 22 to 28 days out of the nest. During the fifth week out of the nest the youngbird nearly attainsadult plumage.By Day 35 the youngmale has only a few smudgybrown feathers on his black andwhite necklace, and the youngfemale can be told from her mothermainly if one looksvery carefullyfor a few dark and smudgyfeathers on the breast. However,the youngat this age are peepingvery freely as they followtheir respective parents about, so onecan tell themfrom their parents. At timesa parentbird supplantsits own young. The youngforage by them- selvesrather well, but readily take food from theirparents; probably the adults givethem over half of the food that they eat. The youngbird is very tame and easyto see;it comesup and peersat the observer,then resumespeeping after the parent. During the sixthweek out of the nest,the youngbird losesthe last smudgy breastfeathers. Except for the slightlyswollen and pale anglesof the gape, the youngbird 42 days out of the nest looks much like the adult. However, it is peepingdesperately as it follows its rather antagonisticparent, which frequentlysupplants it rather than feedsit. Althoughit is gettingmost of its own food, it staysnear its parents. The latter, if the nestingseason has not yet ended, are beginningto seem charmedby each other again. There are male-femalegrooming sessions when the fledglingis as little as 30 daysout of the nest, and at times by 37 days out of the nest the male feedshis mate rather than the youngbird. In one case,a male fed his mate and copulated with her whentheir onlyjuvenile female, in the careof the male,was at Day 29. In anothercase, a male fed his mate and copulatedwith her when their only fledgling,a male in the care of the female,was about Day 20. Once a male tried to hop on the back of his juvenilefemale after a feeding. One male at Cerro Campanatried to do so after feedinga youngmale! If the male has 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 67 been feedinga youngmale or the female a youngfemale, the young bird is quite often supplantedwhen it comesup peepingat the adult'sserpentine- songs.One suchyoung female left her courtingparents and followed a nearby ant swarminto the territory of a neighboringpair at this time. Two other youngleft their parentsabout this time, and the youngfemale nine dayslater was 1,300 m away. More often the seventhweek is the time of the parting of close bonds betweenyoung and their parents. The youngbird foragesfor itself evenwhen it stayswith its parents. It still peepsfrequently, but the parent rarely feeds it. The parent snarlsat and supplantsa youngbird of its own sex. Once a youngfemale snarledback, and her mother promptlysupplanted her several times. The young is sometimestolerated into the eighth week, even in the middleof the breedingseason, but by about56 daysout of the nestall young have left their parents. I was not able to checkthis for younghatched at the end of the breedingseason, however; two broods of these young were still peepingwith their parentson 23 and 25 November 1961, after at least 52 and 54 daysout of the nest. The various kinds of peeping changelittle during the period the young bird is growing,but peepingdoes become more frequentafter the youngbird is flying well. Until then it is very quiet, exceptfor loud-peepingwhen the parent singsat a distanceor disappearsfor a long time. When the parent flies up with food or singsnearby, the youngbird faint-peepsor serpentine- peeps,and it squeaksas it is fed. During the fourth and fifth weeks, as the parent stopsfeeding it, the youngbird peepsa great deal. When the parent is distant,it loud-peeps;otherwise it faint-peeps.At timesthe faint-peeping gradesinto soundslike the adult songwhen the youngbird is five weeksor more in age;one gavesongs after beingbanded. At timesyounger fledglings givehoarse or faint songs,even at Day 21 or so.

RENESTING AND NEST SUCCESS One youngbird becameindependent when somewhatover six weeksout of the nest. The otheryoung bird may have stayedlonger with the parents,but it was independentwhen eightweeks out of the nest. The pair starteda new nest 5 days later, or 17 days after the first young was independent.There may have been an unsuccessfulnest in the interim. Other young became independentat about 50 to 55 days out of the nest, and renestingsstarted up to a week before or after this period. In one case a juvenile female (feedingherself) followedher father and waited,making loud-peeping notes, near the next nest for a week while he incubated,so that renestingstarted at least two weeksbefore independence.There were three other caseswhen a pair had both a nest and a semi-independentfledgling. In one case (the 68 ORNITHOLOGICAL MONOGRAPHS NO. 10 caseof copulationwhen the female's young was only in Day 20), thefledgling wasabout 30 daysold whenthe new nestwas started. Granting35 daysfor a nestingand 30 to 50 daysuntil the resulting fledglings are independentor semi-independentand a new nestis started,it shouldbe possiblefor SpottedAntbirds to rear three broodsbetween 15 April and 15 November. One pair did rear two broods,and had at leastone unsuccessful nestin between. This pair had youngabout 15 daysout of the nest on 3 June 1961 and independentyoung by 10 July 1961. The male was feedinghis mateon 31 July and 1 August,suggesting that oneor moreJuly nestingshad failed. They wereincubating on 7 and 18 August. On 23 Augustthey were feedingyoung in the nest. The next nestingbegan about 30 August,for they had youngabout two daysout of the nest on 5 October. Theseyoung were still with them on 23 November. NormallySpotted Antbirds are luckyto get evenone brood of youngout of the nestper year,for predatorsrob nearlyall nests.Of 24 nestsfound before or duringegg laying,only 6 reachedthe hatchingdate. In 2 of these6 nests,both eggs failed to hatch(both nestswere robbed after more than 16 daysof incubation;the incubationperiod is normally 15 days), and in one nestone egg of two did not hatch(the eggremained in the nestfor threedays, thendisappeared while the growingyoung remained). The other 18 nests(75 percent) were robbedbefore hatching. If the two nestsshowing complete hatchingfailure are countedas failing in the egg stage,20 of 24 nests(83 percent)failed in the eggstage. SinceI left long beforevery smallnestlings in two nestscould leave, young in the third nestdisappeared when between threeand six days old, andonly one of twoyoung left thefourth nest, I have yet to follow a completelysuccessful Spotted Antbird nestfrom egglaying to fledging.In 8 of the 18 robbednests, the eggswere taken duringthe three-dayperiod of egglaying. Two othersets were taken three days after clutchcompletion, one set betweenthree and six days,and one set on the fourteenthday. The other 8 nestswere robbedbetween egg laying and the fourteenth day of incubation. ff the nest is not robbedbefore the younghatch, they still have only a moderatechance to survive. Of 19 nestsfound at or before hatching, one or moreyoung left successfullyfrom 8 (42 percent). If the two abovenests in whicheggs never hatched are counted, 13 of 21 nests(62 percent)were robbed in the "nestling"period. Only one nest,with addledeggs, was lost in the first three daysafter the hatchingtime; otherwiselosses were scatteredir- regularlyduring the nestlingperiod. Of 12 otherbroods I found after they were hatched, 8 were successful.Of 16 successfulnests, 6 had only one nestlingat fledging,and 1 nestthat wasprobably successful added to the 10 to make 11 nestswith two youngat fledging. In all, 16 broodsof youngleft 84 nestsfound at variousstages and followed 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 69 until successcould be determined. Actual nest failure (the percentageof nestsin which one or more eggsare laid but from which no youngfledge) is higherstill, as Mayfield (1961: 255) pointsout; mostof these16 broodswere from nestsfound only late in nesting. Many nestsin which I never found eggsor young(because already destroyed) are not includedin the abovetotal of 84. Probablynest failure is lower than the 95 percentof 22 nestsfound at the start of incubationor before and checkedfor fledgingwould indicate; one nestlingsurviving from each44 eggslaid wouldprobably not perpetuate thisspecies. Nest failure may be lowerthan 93 percent(if m is the proportion of nestslost in the egg stageand n is the proportionof hatchingnests lost in the nestlingstage, the total proportionlost is m q- n [1-m] or 0.83 q- 0.58 [0.17] •0.93). However, nest failure probably averagesover 90 percent. Similarlyhigh rates of nest failure are usual for birds of tropical forests(Willis, 1967: 88; Ricklefs,1969a: 18) and may be usualfor mature forestedhabitats elsewhere(Snow and Snow, 1963: 35). Nest failuresare due to predators,apparently. Snakes are the only known predators.The disturbednest noted abovewas almostrobbed twice by two differentPseustes poecilonota, and waslater robbedby an unknownpredator. I hit with a rottenstick and chasedunderground the first large,hissing, spread- gaped snake and collectedthe secondsmaller one. I doubt that the first one was seriouslydamaged, and the fact that it did not return to the nest before the secondone suggeststhat these snakeswork by wandering and lookingin nestsrather than by revisitingnests found earlier. Since my re- peatedvisits to antbirdnests have not led to greatermortality than at seldom- visitednests, I suspectthat human visits to nestsmay not attract snakeat- tentionunless, perhaps, the snakeis watchingdirectly. Also, that two snakes appearedsuggests that any routes or home rangesmust be wide-rangingor perhapsvagrant or overlapping,as is usualfor snakes. Pseustespoecilonota is a fairly common,big-gaped snake that readily climbsto the tips of small twigs; it may be a nest-robbingspecialist, for the only things I have seen one eat were two nestlingsfrom a nest of Slaty Antshrikes. Since these snakesdo not disturbnests, except by pulling out youngthat graspthe nest linings,and sincesome destroyed Spotted Antbird nestswere rather torn up, or were torn out, or had a hole in the bottom, I suspectother predatorsalso rob antbirdnests. Ricklefs (1969a: 39) has shownthat speciesdiversity of predatorscorrelates roughly with nest failure, and suggeststhat many different speciesrob nestsin tropicalor maturehabitats. There may be someloss of young the first few days after they leave the nest. Of 17 broods at fledging, 6 consistedof only one nestling;but of 58 broods a week or more out of the nest, 26 consistedof only one fledgling. These data, if taken at face value (6:17 = X:58; X -- 20.5; 26 - 20.5 = 5.5), 70 ORNITHOLOGICAL MONOGRAPHS NO. 10 suggestI that 5.5 broodsof 58 (9.5 percent)had lost one of two fledglings after they left the nest. If so, 0.54 pair of 58.5 pairs (0.095 x 0.095 x [58 + y] = y; y = 0.54) had lost both youngby the time I found them. Pos- sibly some6.58 (0.54 + 0.54 + 5.5) youngof 117.08, or as many as 5.6 percent,were lost in the first week after fledging. In 15 or so broodswatched from the first week to independence,no more youngwere lost. Even young with broken legsand one with a club foot (swollen and fused-toed,clublike) survived. If a nestis destroyed,the pair renestsimmediately. In two cases,building the new nest startedabout four days after destructionof a set of eggs. ! was not able to follow singlepairs of SpottedAntbirds throughout a full breeding season.However, no pair has ever beenwithout reproductive activity, either courtshipfeeding or nestingor caring for fledglings,for more than a few days at any time when it was under observationduring the nestingseason. From this, from the knownnesting periods, and ratesof nestdestruction, one can estimatethat individualfemales should average eight or nine setsof eggs betweenmid-April and mid-October. A renesfingrate almost as high as this has been determinedfor ant-tanagers(Willis, 1961: 500) and for Bicolored Antbirds (Willis, 1967: 88).

BREEDING SEASON On Barro Colorado, the breedingseason is the rainy season,from April to November. The earliestnest I found in 1961 had the first egg in it on 18 April. Young approximately16 daysout of the neston 3 Juneprobably came from eggslaid about20 to 22 April. Young left a late neston 23 October;the eggswere probablylaid 26 to 28 September.Young 15 days out of the nest on 22 Novembermust have come from eggslaid about 10 to 12 October. I left Barro Colorado on 26 November, and hence could not determine if there were any later nestings. There is no evidenceof nestingin the period from mid-Novemberto early April. No dependentyoung were evident between 15 November 1960 and 3 June 1961, between 18 January and 18 February 1962, between15 Januaryand 5 March 1965, between18 December1970 and 23 January1971, nor beforeyoung left a nest on 27 May 1966. Adults rarely showedsigns of courtshipbehavior before April in 1961 or in the monthslisted above for 1962, 1965, and 1970-71. In 1961, the numbersof nestslocated by month of egglaying, were, from April to October,respectively: 1, 3, 2, 3, 3, 4, and 1. Young left one of the May nests,one of the Augustnests, and two of the Septembernests. Counting the youngfrom thesefour nests,the numbersof flying broodsrecorded by

• These calculationsprove, as noted in a letter from the author, to be based on in- complete assumptions. The intricate solution must await his return from Brazil.---Ed. 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 71 estimatedmonth of egg laying were 1, 1, 3, 1, 2, 10, and 2 from April to Octoberrespectively. I do not know whethermore youngemerge late in the seasonevery year or not; BicoloredAntbirds were also more successfullate in 1961 (Willis, 1967: 88). Sinceall pairsare nestingmore or lessconstantly from April to mid-November,the greaternumbers of youngout of the nest late in 1961 indicate that in 1961 early nestsfailed more consistentlythan did late nests.

DISCUSSION SpottedAntbirds, like BicoloredAntbirds but unlike someother species (see Nice, 1943: 174; Akerman, 1966a: 326; Moynihan, 1955: 86; Wood- Gush, 1956:141 ), useagonistic behavior very little in courtship.Preliminary challengingof a potential mate, probably not instantly recognizedas an eligiblemember of the oppositesex, is so brief in the antbirdsthat it is scarcelymore than an initial repellingreaction, rather than a basic part of courtshipas are threateningbow-coos in pigeonsor "pouncing"in Song Sparrows.The studiesof suchspecies as chickensand pigeonshave given rise to the widely but usuallyuncritically accepted "FAM" hypothesis,that courtshipresults from conflictingtendencies to flee, attack and mate (see Hinde, 1969, and Bastock, 1967, for generalreviews). SpottedAntbirds form permanentpairs only on territory, which suggests that singing,dominance, and other signsof territorial statusmay be necessary for preliminary stagesof courtship. Clearly, agonisticbehavior can help each bird get its territory and keep it, drive away competitorsfor food or for the mate,prevent polygamy and biologicallyunproductive indulgence and rape (Marler and Hamilton, 1966: 191), and help in many other ways necessaryfor life and reproductionbut peripheralto the actual courtship displays.However, in SpottedAntbirds singing is the only part of courtship that could be consideredagonistic behavior. This raises the questionof whethersinging to mates or young is to someextent "hostile"or agonistic behavior. Perhapspartly as a reactionagainst the excessesof early naturalists,who as- sumedthat birds sing only for happiness,scientists sometimes have seemed recentlyto presumethat birds sing out of surliness.Early naturalistsmay have been partly right, in the sensethat "pleasurecenters" in the brain probablyare part of the physiologicalsystem controlling singing (if singing is necessaryfor survival, such pleasurecenters are very likely to evolve); and later scientistsmay be rightin the sensethat any pleasurecenters probably evolvedbecause singing is an efficient and safe way to ward off rivals, keep a family togetherin denseundergrowth, attract a potentialmate, etc. There is little doubt that singing,especially loud singingassociated with calls and movementsof the challengingdisplay, helps set up territoriesand causesflight 72 ORNITHOLOGICAL MONOGRAPHS NO. 10 or fight reactionsin rival Spotted Antbirds. There is also evidencethat singing,especially faint songsassociated with chirping,helps keep mates and youngtogether. For many otherbirds singing is an activitythat fits both in a chapteron agonisticbehavior and in one on reproductivebehavior. (Moyni- han [1962b: 319] haseven proposed that songbe definedas vocalizationsthat usuallyattract conspecific birds of the oppositesex and repel onesof the samesex; but this dependson the situation,for commonlysongs attract a hot rival of the samesex and repelmembers of the oppositesex. I suggesta con- ventionaldefinition for "song"instead: the so-named(by the first students) major vocalizationsof a species,especially its longestor loudestor most musicalor far-carryingones.) Althoughsome may preferto invokeOccam's Razor to assertthat one motivation--sayhostility--should stimulate singing or similarvocalizations rather than differentmotivations in different contexts, I prefernot to assignhostile motivation to courtshipsong in SpottedAntbirds becausechallenging seldom is associatedwith it. Moreover, if it is hostile, the male, who singsless once he getsa mate, and singsmore if temporarily separatedfrom her, must be expressinghis hostilityto her until she arrives! (I am indebtedto Lack [1943] for disposingof the possibilitythat birdssing out of happinessby usingsimilar observations.) It is more likely that song is a long-distancecommunication of deprivationof a mate, an assertionof presence,etc.; thosewho invoke hostilityshould prove it existsin the given situation. If courtshipsong is not likely to be hostilein SpottedAntbirds their courtship is essentiallyan "M" type (Bastock,1967: 102), in contrastto the "FAM" type of somebirds. FAM birdstend toward M displayslate in courtship, and occasionalelements of aggressionand submissionappear even in SpottedAntbirds; there is a differencein degreerather than in kind. The very commondisplay or activityof courtshipfeeding may have evolved from a flight-fightconflict or from a foraging-matingconflict, but seemsmore likely to have comefrom parentstransferring attention from fledglingsto matesat the start of new nestings.If so, any conflictwas probablyirrelevant and betweenparental and mating drives. Birds sometimesbring food to a nestbefore young hatch, but suchinstances seem unlikely to leadto courtship feeding.Courtship feeding generally precedes egg layingexcept in species in whichfemales incubate alone but are fed by males. The ecologicalorigins of thesecourtship displays need studyas much as do thepsychological origins. Courtship displays require expenditure of energy and time and put the displayerin somedanger, especially if specialcolors or structureshave evolvedto emphasizedisplays. In SpottedAntbirds and other courtship-feedingspecies, the male losesfood when he givesit to the female. Not only mustsuch disadvantages be offset,there must be suchan advantageto the displayersthat their geneseventually replace those of non- 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 73 displayers.The mostlikely advantagesto displayersare that they can evolve sexualisolation from relatedspecies and be sexuallyselected by or stimulate potential mates. If thereare other specieswith whicha bird can hybridize,and the offspring of such unions survive poorly, the female and to a lesser extent the male must develop the ability to tell its own species. Differencesfrom related specieswill be selectedfor, whetherdifferences in plumage(Sibley, 1957: 174) or displays. Displays are probably easier to select for, since changesin morphologyinevitably compromiseother adaptations,such as those for avoidingpredators or for gettingfood, while displayscan be turned on and off. If the specieslearns the proper displaysfor a mate, and perhapseven if a changein the neural system(instinct) is required, necessarychanges can perhapsbe mademuch faster or more easilythan can changesin morphology. Possiblythe lack of striking courtshipdisplays in SpottedAntbirds, which have no speciessimilar with them, is evidencefor the hypothesisthat species withoutclose relatives are unlikely to developstrong courtship displays. The hypothesisneeds checking, although Sibley explainsthe loss of bright male plumagesin duckswithout sympatricspecies in this way. Another factor in the evolutionof displays,however, must be the stimulation, attraction, and synchronizationof a female and male by displays. Lehrman(1958: 33) hasshown experimentally (using injections of hormones) that displaysin Ring Doves (Streptopeliarisoria) stimulateand synchronize the mates. Stimulationand attraction are probably important when mates meet briefly, have a shortbreeding season, or are limited in availablehabitat. The shorterthe time or the smaller the spacein which male and female can be together,the strongerthe displaysor songsor colorationof the male shouldbe in order to bring the femaleinto readinessto mate in a short time, to attracther to him rather than to anothernearby male, and for her to find him in an isolatedpatch of habitat. (Similar argumentsapply for the displays or songor colorationof the femalein suchbirds as jacanasand phalaropes.) The needfor synchronizationshould also lead to strongcourtship displays in seasonalor isolatedhabitats; strong displays,even communalones, will quicklybring birds into breedingcondition. The courtshippatterns of birds may thus be influencedby the habitats theylive in. Verner and Willson ( 1966: 145) havepointed out that polygamy, with its consequentemphasis on persistentmale displaying,arises mainly among marsh and open-countrybirds in North America. They theorize that the great variation in quality of habitat from one site to another in im- matureand low habitatsmakes it possiblefor somemales to preemptsuch good territoriesthat severalfemales can raise more young in themthan they could by spacingthemselves out with unmatedmales on poorer sites. The Spotted Antbird and its relativesrepresent the otherextreme, monogamous birds with 74 ORNITHOLOGICAL MONOGRAPHS NO. l0 prolongedpair bondsand equal roles of the sexesin caringfor offspring, characteristicof forest or mature habitats. In the tropics,the birds that use this system(antbirds, ovenbirds,trogons, puffbirds, etc.) predominatein maturehabitats while thosethat displaypromiscuity or female care at the nest are often predominantin second-growthor "edge"habitats( tyrannids, icterids,tanagers, etc.) or feed on irregularlydistributed nectar or fruits (manakins,hummingbirds, etc.). Parrots and woodpeckersare major exceptionsto this rule, since both have irregularlydistributed habitats and monogamoussystems; but mostare large, far-flying birds that can easilyand safelymove from one patch to the next and thus are not in an environment that is patchyconsidering their powersof flight. Differencesin pairingsystems between forest and savannabirds are known from the samefamily and genus.In the weaverbirds,genus Ploceus, the forest speciestend to be territorial and monogamous,and male and female are bright-coloredand incubate. Speciesof savannatend to be stronglydimorphic and showpolygamous and colonialbreeding in which the female incubates aloneand the maledisplays conspicuously (Moreau, 1960: 321; Crook, 1964: 104; Lack, 1968: 36). In grouse(Tetraonidae) speciesof widespreadmature habitats(tundra or woodland)tend to be monogamousor solitary-polygynous and have moderatedisplays, while speciesof localizedhabitats (edges of woodlandmeadows) and grasslandtend to be communal,promiscuous, and strongdisplayers (Scott, 1963: 673 ). Orian.s (1969a) has recently suggestedthat second-growthand marsh or grasslandplants put a large part of their energyinto fast growth and quick dispersalof many seedsrather than into biochemicaland other protective devicesagainst insects, hence tend to be infestedwith insectsonce the insects find them; therefore,second-growth plants have at any one time someabun- danfly fruiting patcheswhich insectshave not yet found and othersthat are stronglyinfested. Moreover, greater light intensitiesin the canopy and in secondgrowth probably lead to great productivities(Pearson, 1971: 50). Either the reproductiveparts or the insectsare locally superabundant,and producelarge local food suppliesfor birds exploitingthem. Superabundantfood suppliesboth releasethe male from feedingthe young or helping the female at the nest and permit some males to preempt local good sitesand get as many femalesas possible. Strongdisplays and songs are amongthe secondarysexual characters that such males can use to drive off other males and attract more than one female each, and the males that do so will leave more offspring. Thus, the FAM mating system,in which the female is impressedby a strongmale and there have to be quick and distinctive displays,is perhapsmore likely to evolve in ephemeralor seasonal habitatsthan in climax and constanthabitats. In the latter there are many species,each exploiting a sparseand relativelyevenly distributed food supply, 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 75 and no one areais so muchbetter than anotherthat one male can preemptit or havemore females because they do betterto go to him than to his poorer bachelorneighbors. Monogamoussystems, especially ones in whichthe male helpsthe female at all stagesof nesting,will probablybe mostsuccessful in matureand widely distributedhabitats. Monogamoussystems make it possiblefor male and femaleto shareduties of caringfor eggsand young,thus partly compensating for a habitatwithout local food surpluses. The monogamoussystem does not requireconspicuous courtship displays, especially if there are no sympatric closelyrelated species. Sympatric congeners are mostcharacteristic of genera of secondaryhabitats or patchyhabitats, probably because irregularity in time or spaceallows "fugitive" species(Hutchinson, 1951) to skip about aheadof more efficientcongeners as well as favoringthe good powers of flight or dispersalmechanisms that make skippingabout possible.In mature (homogeneous)habitats, sympatric congeneric species almost always separateout by smalldifferences in habitat or quicklydevelop morphological differencesthat causeus to classifythem in differentgenera. The big genera in the tropicsare often ones predominantlyof secondgrowth and patchy habitats,and showtendencies toward strong displays or care of the nestby the female alone. In SpottedAntbirds, an importantadvantage of courtshipfeeding may be thatit helpsthe female form the eggsand avoid predation during a difficult period. The rarity of courtshipfeeding at times when the female is not formingeggs, with the exceptionof the pairing period, suggeststhat egg- layingand courtshipfeeding may be linked. Courtshipfeeding may alsoin- dicateto thebirds whether it is possibleto feedyoung birds, which will appear in a few weeks(if it is possibleto feed a female,it shouldbe possibleto feed young);but thereis probablyso muchday-to-day variability depending on whethera swarmof antsis presentor not that only generalindications would be possible.The two-dayperiod between laying eggs suggests that the female mayfind it demandingto getfood to form the ratherlarge eggs. Presumably smallereggs could be laid, but at the expenseof requiringmore food for the young later on and more visits to the nest (Lack, 1968: 190). More visits,besides possibly being more dangerous to theyoung (Skutch, 1949: 434) probablysignificantly increase the dangerof predationon adults. The sharp chippingnotes of SpottedAntbirds visiting nests suggeststhat they are alarmed, probablyjustifiably so when one considersthat Semiplumbeous Hawksand other sit-and-wait predators would find it easyto captureSpotted Antbirdsmainly when they are on the move. The higherrate of feedingnest- lingswhen pairs staynear them may be possibleboth becausepredators are lessdangerous on shortmoves and because the birdsdo not haveto spendso much time travelling. 76 ORNITHOLOGICAL MONOGRAPHS NO. 10

There is no obviousreason why matedSpotted Antbirds groom each other while related antbirds,such as the Bicoloredand OcellatedAntbirds, do not. The latter specieskeep their mates for life with just flirting and courtship feeding. SpottedAntbirds tolerate the mate at very dose quarters,perhaps partlybecause the maleand female differ in colorand do not provideagonistic stimulifor eachother. Perhapsa small antbird,eating common small insects and ordinarilyforaging some distance from the mate, can tolerate the mate readilyon the few occasionswhen they near one another,thus predisposing for grooming. However, the OcellatedAntbird toleratesthe mate and does not groom. It is possiblethat SpottedAntbirds more readily pick up head parasitesbecause of their frequentforaging above the groundand through vegetation,hence need groomingmore. I have seenmore ticks on Spotted Antbirds than on Bicolored and Ocellated Antbirds, but would not vouch for the statisticalsignificance of such rare events. However, I have little doubt but that the evenrarer eventsof predationcause striking behavior patternsin SpottedAntbirds, so that dangeror irritationof headparasites or spider webscould easily cause grooming to be favoredin evolution. The Spotted Antbird probably has little trouble finding its simple re- quirementsfor nestsites, but the visitingof old nestsand the occasionaluse of the same nest site by the same or different birds suggestthat there may be somelimitations. The low heightprobably reduces competition for sites with the lowestnesting individuals of the SlatyAntshrike, as doesthe tendency to nestclose to stemsrather than away. However,Spotted Antbirds nest in the nextto highestquarter of theirusual foraging range and SlatyAntshrikes in the bottomhalf of theirusual foraging range, suggesting that pressuresother than interspecificcompetition are more important--possiblypredation or lack of twigsat low and high levelsin the forestunderstory. The generalconstruction of the SpottedAntbird nest is loose but fairly opaque,probably so the nest resembles trash, while the nest of SlatyAntshrikes is nearlywoven but thin and translucent.The low-nestingRed-throated Ant- Tanager,with a trashynest, and the high-nestingRed-crowned Ant-Tanager, with a thin nest,show an evenmore strikingcorrespondence between nest type and height of nest (Willis, 1961: 482) than do the antbird and antshrike. Low in the undergrowththe trashlikenests blend with leaveson the ground from the viewpointof an aerial predatorand with pilesof debrisfrom that of groundpredator. High in the undergrowth,trashlike nests would be sil- houettedagainst green leaves to the eyesof both aerial and groundpredators, and the former would see the eggs,so that flycatcherswith trashlikenests tend to haveroofed nests and to put them at the tips of barely accessiblevines or twigs;birds with open nestsmake them thin or decoratethem with green moss so green will show through from both directionsand the pale eggs will be lessclearly silhouetted from above. 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 77

Possiblynests of SpottedAntbirds are not placed in patchesof dense undergrowthbecause some predatorson nestsor adults could approach unseenor learnto work suchpatches. Placing nests in scatteredpatches rather than in the uniform openforest matrix makesit easierfor predatorsto locate the nestsif predatorsspecialize. Also, the greenleaves of suchpatches could frame the nest for aerial predators that would overlook it against dark backgrounds. At the time of building,the birdslook for materialnearby and moveto the nestrather conspicuously and frequently.The femaleis more disturbedby an observerthan is the male, but both chirp and singand react more strongly to alarm callsof treetopbirds than to the observer.Presumably there are no animals,such as brood parasitesor predatorswith long memories,that could causeselection for more cautionon the part of the parents;their caution seemsmostly for themselves.Looking for material near the nest reduces the energyneeded for buildingas well as dangersfrom predation,as does concentratingbuilding rather than scatteringvisits out. Whether lookingfor material at unevenlydistant sites and irregulartimes is better amongbirds subjectedto brood parasitismis not clear, althoughone could suggestthese or otherreasons for irregularsearching and buildingin the numerousspecies in which it occurs. In any case,the SpottedAntbird is not known to be parasitizedby cowbirdsor otherbrood parasites. The great amount of chirping and flirting during nest building probably helpssynchronize the mates,and perhapshelps start hormonalpreparations for egg-layingby the female and for incubation,in which both birds must participate. Later thereis muchless interaction at the nestbecause then eggs and youngmight be detectedif the pair were conspicuous. Laying the eggsan hour or two after dawn seemscharacteristic of antbirds, as opposedto ant-tanagers,which lay eggsat dawn (Willis, 1961: 485). It seemsstrange that a female would carry a large egg about for a couple of hours, unlessshe needsfood before she can lay or unlessit is dangerousto go to the nestwhen it is still dark in the forest. However, Skutch (1952: 55) notesthat flycatchersof open countryalso lay eggswell after dawn; laying at dawn seemscharacteristic of some songbirdsand hummingbirds,laying later characteristicof shrikes(Susan Smith, pers. comm.) and of suboscines. Is later laying perhaps an adaptation to generally sparsefood supply? to difficultiesin storingup enoughfood the eveningbefore? Seed and fruit eatersoften can storefood up the eveningbefore. The heavyblotching of eggsof SpottedAntbirds probably helps them blend againstthe backgroundof deadleaves when seenby an aerial predator. Slaty Antshrikeshave almostwhite eggs;but sucheggs are not so conspicuousin a thin nest, and parentsincubate them nearly all the time (Yoshika Oniki, pers. comm.), while Spotted Antbirds sometimesleave the nest unguarded 78 ORNITHOLOGICAL MONOGRAPHS NO. 10 before the mate arrives. (White-flanked Antwrens, which have white eggs in low and thick cup-shapednests, usually put the nestbelow an overhanging leaf; they sometimesleave the nestunguarded.) PossiblySlaty Antshrikes,which are not ant-followers,can be on the nest constantlybecause they forage near the nest, while SpottedAntbirds often forage a long distanceaway and at swarmsof ants. In such cases,the in- cubatingbird shouldstay as long as it wantsrather than beinginterrupted so as to causemore long trips. Moreover, the bird on the nest can find an ant swarmthat its mate has discoveredbetter by goingto the mate off the nest rather than by random searchafter the mate has come to the nest. Bicolored and OcellatedAntbirds also commonlyleave the nest to go to the mate at swarms. SpottedAntbirds visit the nest slightly more frequently during the day than do their larger relatives,the BicoloredAntbirds; perhaps a smallerbird metabolizesits food storesmore rapidly than does a larger one with similar food habits. Skutch (1962: 134) noted how long the incubationsessions of antbirdsare consideringtheir size,and suggestedthat longerperiods might result in hunger. Incubation by both sexes allows individuals of this and other antbirdsto incubateless than 50 percentof the daylighthours without feedingof incubatingmates. Incubationby the femalealone is rare in forest- adaptedinsectivorous birds in the tropics(Willis, 1969b), probablybecause feedingthe female or leaving the eggsunguarded would attract predators, while the female could not get enoughof her scatteredand uncommonfood if the male did not help her. Carrying the eggshellsaway when the young hatch, rather than eating them as do ant-tanagers(Willis, 1961: 488) or letting them be crushedas do somehummingbirds (Oniki, 1970: 724) is a minor activitybut probablyhas somereason (Tinbergen, 1963: 28). Perhapshummingbirds, with their thin bills, cannotcarry shellsaway or eat them, and as a consequencehave soft whitishnests that alloweggs to be thin-shelledand crush readily after hatching; white eggshellswould be conspicuousin a dark nest,such as that of antbirds. It could be that antbird eggshellsare too thick or large or too coveredwith pigmentsto be eaten, hencehave to be carried away. Antbirds with nearly whiteeggs (Slaty Antshrikes,according to Y. Oniki, pers. comm.) carry the shellsaway, suggestingthat the pigmentsare not the problem. Antbirdsusually dissectinsects of the size of eggshellsafter takingthem to the ground,suggestingthat they do not have the tongue-billcooperation or the breadth of bill that allowssequential crushing of an item (as in ant-tanagerseating fruit, Willis, 1960a: 163) without slingingit about conspicuously.Perhaps the calciumthey couldget out of the shellsis not scarceenough to be worth attracting the attention of predators by conspicuousshell dissection. This suggeststhat thick-billedbirds nesting in opennests should generally crush 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 79 and eat shells,medium-billed ones shouldcarry them away, and thin-billed ones shouldhave soft white-linednests and crushableeggshells that are left in the nest. However, the soft white nest-liningsof goldfinches(Spinus spp.) probably have some other basis, although these are small-billed finches specializedfor small seeds.Also, shorebirdswith rather thin bills commonly carry off eggshells(Sutton, 1968: 502-503). Chirpingand serpentine-singingare associatedwith feedingthe young as well as with feedingthe mate;they seemto be generalattractive notes in this species.Perhaps it is more correctto speakof flirting as an attractingactivity rather than as a courtshipactivity. Of course,courtship feeding itself is like feedingthe young, but servescourtship functions; hence it does seemlogical to classifyit and flirting and groomingas courtshipactivities even though they have other functionsor are like unrelatedactivities. Young are not broodedafter the first few days. This releasesthe parents to forage more and feed the young twice as often at a time when the young need more food, but dependson how fast the young can gain temperature regulation. R. Ricklefs (pers. comm.) notes that putting food energy into early temperatureregulation (as down feathers) takes energy away from growth,hence birds that broodtheir younghave faster-growing young. Spotted Antbirds have downlessyoung that grow very rapidly, probably at almost as high a rate as is possibleconsistent with developingtemperature regulation soboth parentscan foragefor them;predation rates probably favor fast growth of young. The loweringof growthin weightthe lastfew daysin the nestmay representchanneling of energyinto feathergrowth so young can leavethe nest and thus avoid predators,but may just reflectloss of water (Ricklefs, 1968: 34) with feather growth. The high feedingrate of nestlingsin Days 8 and 9 is probablyfacilitated by the parents'foraging close to the nest. They thus reduce travel time. Danger to themselvesfrom predatorsis also reduced,for the adults can watch for predatorsnear the nest while foraging. Undetectedpredators could find the nest more easily, however, becauseof the short parental trips and high feedingrates. Foragingaway from the nest probablyfavors few trips with rare large food items; foraging near the nest favors bringing common smallitems. The youngperhaps do not get as much extra food as the increase in feedingrate would suggest,but may get as much food or more than in previousdays. The local populationof arthropodsnear the nest is probably reduced,and this may explainwhy feedingrates slow on Days 10 and 11 even thoughparents continue to foragenear the nest. At any rate, it wouldprobably not be possibleto forage near the nest for the full nesting period without reducinglocal prey populationsor attractingthe attentionof predators. Foragingnear the nestmay be an advantageduring the last daysthe young are in the nestfor anotherreason: youngold enoughto jump out of the nest 80 ORNITHOLOGICAL MONOGRAPHS NO. 10 can be rushedout of the nestand led away by the parentsif they seea predator approaching.If this is the main reason,the increasein feedingrate may be unimportant. Probablythe increasein nest-leavingcalls and activity by the youngkeeps parents close to them. The suddenincrease in feedingrate when one of two younghas left the nest,usually due to the suddenincrease of feedingby the female,is also interesting.It may be that the low feedingrate of Days 10 and 11 is a result of warinessabout predators,for one parent gets food near the nest as fast as two the momentthat one younghas left the nest. The youngthat hasleft the nestis soonoff in a safersite with an unexploitedfood supply,of course. One beginsto suspectthat the low feeding rate of Days 10 and 1 l may encourageyoung to leavethe nestor be necessarybecause of dangerfrom predators,rather than be causedby lack of food near the nest. At one nest the femalesoon started feeding two youngrapidly on Day 9 when the male, scaredaway by the observerat the start of observations,deserted them for overan hour. It may be, however,that the femalenormally forages moderately far from the nestat thistime, bringinglarge prey, and thusis merelyavoiding competitionwith the male rather than withholdingfood from the young. Without exactknowledge of weightsof food and of distancesat which the adultsare foraging,plus more data to make surethat thesechanges in feeding rate are normal, one can say little more. Skutch (1949) suggestedthat tropicalbirds are not feedingtheir youngas fast as they can but insteadare avoidinggoing to the nest so rapidly as to attractpredators, while Lack (in an addendumto Skutch'sarticle) reiteratedhis earlier (1947-48) suggestion that birdsin generalfeed their youngas fast as they can. The suddenchanges in foragingrate by SpottedAntbirds favor Skutch'sview, for a doublingof foragingrate is not possiblein a systemin which youngare beingfed at a truly maximal rate. However,Lack's generalview is correctif SpottedAnt- birdsare justforaging temporarily near the nestat ratesthey couldnot sustain for longwithout lowering local food supplies drastically. Moreover, the weights of food going to young probablydo not differ as much as do the rates of feeding. One cannotyet decidefor or againstSkutch's or Lack's theseson the basisof data from SpottedAntbirds, but it would be interestingto know if feedingrates are maximal or are slowedby the danger of predation. I suspectthey are nearlymaximal given the variousdangers of predation--i.e., SpottedAntbirds could forage closerto the nest on the averagewere it not for predatorson adultsand young,but are feedingthe youngabout as rapidly as is safe. Probably predationpressures from snakesand other nest-robbersmake it adaptivefor SpottedAntbirds to leave the nest on Day 11, when they can barely fly. Suchyoung are separated,one with each parent, so a predatoris unlikely to get both young at once as it could in the nest. Rates of loss of 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 81 fledglingsare much lower, even at this stage,than rates of lossesof nestlings. Also the parentcan lead the youngto a new placeif predatorsare searching nearby. Later, when the youngfly better, the family reunites. R. Ricklefs (pers. comm.) has pointed out that in birds the amount of parental alarm and attack is correlatedwith how much they have to lose; bird specieswith few eggstend to defendor displayless than birds with more eggs,and there is an increasein defendingas the eggsor young develop,up to the point where young are flying well. In other words, the greater the investmentof parental energy, the greater is the importance of defending this investment. At their nests Spotted Antbirds perhaps tend to flee and chirr becausethey have only two eggsor young and would run too much dangerto themselvesor to the nestwithout concomitant gain in the probability of youngsurviving. Once the youngis out of the nest,however, the parenthas expendedmuch energyon it and also can give the young a fair chanceof escapingpredation by simpledisplays that do not put the parent in much danger. The limited displaysof one female near a nest with eggs suggestfrozen motions of conflictingbehavior patterns,perhaps between incubatingand attackingor fleeing. The flitting, flicking, brief intervalsof challengingor mobbing,and other agitatedactivities of a parent when one is near a young bird also suggestconflicting tendencies. There are ethologlsts(Armstrong, 1949) who feel that distractiondisplays generally arise from suchconflicting tendencies,as well they may evolutionarily. However, the strongdistraction displaysin SpottedAntbirds are so distinct (the flutteringrun, the shielding flight, the shot-birddisplay, etc.) and ritualized that they seem quick and direct reactionsto a situationrather than the hesitant,irrelevant and alternating activitiesone commonlygets in conflict situations.There shouldbe selection for directreactions in situationsof greatdanger, not selectionfor finely graded or irregular responses,unless an irregular responseconfuses the predator. Only neurophysiologicalevidence will showwhether such displays are caused by conflictingneural inputs. FledglingSpotted Antbirds showabout the sameweekly featheringchanges asdo the largerBicolored Antbirds, even though one mightexpect a smallbird to grow and molt and becomeindependent more rapidly than a larger one. PerhapsBicolored Antbirds, which generallyfollow the rich food suppliesat swarmsof ants and are moderatelyhigh in the peck order, are better able to supporta rapid growth rate once the young are out of the nest than are the SpottedAntbirds, which are subordinateand must go through occasional periodswith relativelyless food (see the sectionon foragingbehavior). Lack (1968: 289) hassuggested that slowgrowth rates in nestlingswifts and others are adaptationsfor periodic food scarcity,partly becausea slow-growing bird can put more of its energyinto fat reserves. 82 ORNITHOLOGICAL MONOGRAPHS NO. 10

SpottedAntbirds renestrapidly if eggsor young are destroyed,but wait almosttwo monthsbetween nests if young are raised to independencefrom the first nest. R. Ricklefs (1969b) has found that birds of northern lands commonlyrenest nearly as quicklywhen youngare fledgedas when they are destroyed,while tropicalbirds commonlywait longerif young are fledged. He interpretsthis as a reactionto predationpressures and food availability for youngand time availablefor breeding:in the tropics,with high predation on nestsand less restrictionon breedingseasons, each young that fledgesis valuableand worth a long period of care; in the north, where there is less predationand young can find food with little difficulty,they gainindependence while still in juvenalplumage so that parentscan raise anotherbrood before the shortbreeding season ends. In Spottedand Bicoloredantbirds, the dependentperiod comesto an end as the young gain their first adult plumage and adults become aggressive towardthem, while in northernbirds the youngseem more oftento be forsaken despitetheir juvenalplumage. Perhaps for this reason,northern ornithologists have not emphasizedthe value of the juvenal characters as nonaggressive stimuli (exceptHinde, 1961: 405). Ethologistsand psychologistshave suggestedfor yearsthat the big eyesand headsof "babies"elicit nonaggressive behavior in humans. However, these characters have other functions and probablyare not evolvedspecifically to lower parental aggressiveness. Although such tropical birds as SpottedAntbirds have low clutch sizes comparedwith birds of higher latitudes,they generallyproduce many eggs per year becausethey renest repeatedly and havelong breeding seasons. Prob- ably many pairs manage to produce a few young each year, despite low clutchsizes and extremelyhigh ratesof nestpredation. The Spotted Antbird stops nesting a month earlier than the Bicolored Antbird, which nests from early April to mid-December. The month of November is the wettest month on Barro Colorado, with an annual average rainfall of 454 mm. Possiblythere is too muchrain in Novemberfor Spotted Antbirdsto incubateor feedthe youngand still get enoughfood for themselves. Small birds have a proportionatelymore rapid metabolismthan larger ones, and SpottedAntbirds exchangeincubation duties and feed the young more often than do BicoloredAntbirds. If excessiverainfall interfereswith Spotted Antbirds'getting food, they shouldbe unableto nestand hencerare in regions where rainfall is over 400 mm per month for many months of the year unless the rain falls mainly as occasionaldeluges or at night. Slud (1964: 221) found fewer in rainy parts of Costa Rica than in a periodicallydry region. I found a few at Tanand6 and Yuto, up the Atrato River from Quibd6 in westernColombia, in a regionof about 500 to 900 mm rain per month. The specieswas absentat similarlyrainy E1 Tigre (4 ø 57' N lat., 76 ø 30' W long.), whereI spenttwo weeks. It is muchmore commonto the north, in Antioquia 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 83 and C6rdoba,in areasof lessrainfall. However,the rainfall may be causing an importantchange in food supplyor forest structurerather than interfering directlywith the nesting.Usually there are more kinds of professionalant- followers in wet regions,for instance. It is morelikely that the SpottedAntbird stopsnesting earlier than doesthe BicoloredAntbird becausethe former is low on the interspecificpeck order at swarmsof ants, or becauseits food supply tends to fail earlier. In the former case,migrant thrushes at the swarmsof antscould be interferingwith the nestingof SpottedAntbirds. Although I could easily have overlooked young SpottedAntbirds the first October and November (1960) I was on Barro Colorado,it is alsopossible that the hugeinflux of thrushesand other migrantsthat fall preventedthe SpottedAntbirds from raisingyoung successfully. FledglingSpotted Antbirds were certainlynumerous the next October and November,in a year when migrantswere poorly representedat swarms. However,the 1961 and 1965 recordsfor SpottedAntbirds indicate strong nestingin Septemberand October,even though the latter is the peak month for migrants at swarms of ants. The failure of food supplyseems a more likely reasonfor the termination of SpottedAntbird nestingin early November,before the end of the rainy season.Samples of arthropodsfrom the leaf litter on Barro Colorado,taken biweeklyin 1961 from late February to mid-November,show a peak of sowbugsearly in Juneand a progressivedecline into November. The general pattern of total arthropodnumbers is similar. Apparently leaves fall and accumulateon the forestfloor duringthe dry season,leading to a springand summer"bloom" of arthropodsthat consumemost of the litter by November (Willis, MS). Sowbugsare an importantprey item for SpottedAntbirds. The larger antbirdstake large prey and such predatory arthropodsas spiders, which hide under limbs and treefalls and should be most abundant after the peak of abundancefor small prey. However, much more information is needed,especially data on correlationsbetween nesting seasons of Spotted Antbirdsand rainfall and insectabundance in otherregions.

WANDERING YOUNG AND TERRITORIAL ADULTS: SPATIAL BEHAVIOR Once a youngSpotted Antbird leavesthe territory of its parents,it wanders alone exceptfor associatingwith birds of other speciesin forest flocks and over swarmsof ants. Adult SpottedAntbirds chase it away,and it ignoresor evadesor chasesaway other wanderingimmatures of its own species.Both male and female immatureswander more or less nomadicallyuntil they are several months old. In foraging, nomadic immaturesare clumsy for a month or two after leaving their parents. The immatures seem to expend much energy flying 84 ORNITHOLOGICAL MONOGRAPHS NO. 10 around,chasing prey, and evenpicking up bits of leavesor twigs. At times they are forcedto expendenergy because dominant birds restrict them to poor regionsat the peripheryof ant swarms. However, a wanderingimmature that is away from ants or away from competitorsover ants often seemshyper- active comparedwith older birds in suchsituations. Nomadicimmatures are very quiet;they chip and chirr much lessreadily than do adults. Their chipsand chirrs are often falsettofor a month or so. The immaturesflee readilyfrom the observeror from a hawk, but seemto be poor or reluctantat performingpanicking or mobbingdisplays. In the mist net, an immaturechips or whimpersvery infrequentlyand weaklycompared with an adult. The older an immatureis, or the lessoften it has seenthe observer,the more likely it is to chirr at him or to flee in a chippingpanic ratherthan to flee silentlyor to becometame or to investigate. Agonisticbehavior is alsoinfrequent among young birds. They usuallydo not sing, so that territorial adultsseldom detect them unlessthey come to the songsof the adults. When detected,the youngbird usuallyevades the challengesand supplantingsof an aggressiveadult rather than showingchallenging and cringing behavior. At times one loud-peepsor faint-peepswhen supplanted. Immaturesover six monthsold perform cringingmore often, and a few youngerfemales begged with flutteringwings as if to a parent. More commonly,nomadic immatures perform displacementbehavior, flitting the wingsor flickingthe tail, toe-looking,bill-wiping, etc. There is seldomany aggressivedisplay. Usually an older immaturesupplants a youngerone with little display. At times young males show that they can perform clumsy challenging displays. Once the dependentfive- and six-week old young males of two families,wandering around the sameswarm of ants,spread into low-challenging posesat eachother. They mixedthe postureswith flitting the wings,flicking the tail, and chippingas if highlyexcited or panicking,and did not snarl. Soon they followed their parents away. Another young male showedadult sequencesof agonisticbehavior when only 90 days out of the nest, or about 116 days after egg laying. He had barely been driven away from his parentsat 58 days out of the nest and been driven away by the male of the adjacentterritory at 64 days, when he enteredthe next-to-adjacentterritory at about the time a hawk killed the residentmale. (Following the trail of an antswarm,I found on the leaf litter a spotof blood and a few of his feathers,plus a white featherof a Semiplum- beousHawk or, less probably, a White Hawk). The residentfemale fed both their youngbirds, about 30 days out of the nest. Four days later the wanderingyoung male helpedher in a ferociousdispute with a neighboring pair. He used high- and low-challengingas noisily and persistentlyas any mated territorial male, but was somewhatclumsy in his displaysand gave 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 85 86 ORNITHOLOGICAL MONOGRAPHS NO. 10 rather falsetto snarls. He was still paired with the same female and on the same territory one year later. No other young male showedsuch agonistic displays.Presumably this was becauseno other male studiedgained a mate and territoryin the breedingseason when he was hatched. Normally, a young male wandersfor a few monthsand then settlesin a broad area overlappingthe cornersof severaladult territories(Figure 19). He may shiftto a completelynew area or shift his homerange gradually, but he does not wander nomadically. Ordinarily he starts singingand stops evadinglocal territorial males in the rainyor breedingseason after he hatched. At timeshe startsin the dry season,only a few monthsafter hatching. The area is sometimesone he has occupiedthe precedingseveral months, some- timesa new area. He singsfrequently if he doesnot have a mate. He disputes with neighboringmated malesbut tendsto be pushedtoward the center of his area by them, sometimesto the extentthat he desertsthe area and com- mencessinging in a lesscrowded one. In his areathe youngmale is fairly noisy at chippingand chirringas well as at singingand snarling.Displaced males, like wanderingimmatures, are not noisyuntil they find new areas. Usually the youngmale finds a mate by or duringthe first breedingseason after he hatched,and older malesare unmatedonly when they have lost matesto a neighboror whentheir matesdisappear. Once a male is mated and startsnesting on an area, he may shift it gradually over the years, but he seldom desertsthe area for a new one. So far only one male out of over 100 has shiftedhis territory acrossan intervening territorywhen his mate disappeared.Figure 19 showsmore normal patterns of shiftingwith the years. Male YBXR, followingthe disappearanceof one mate, stole the mate of his neighborto the southwestand graduallyshifted into the area shehad occupied(1963-64). When he lost this mate, he wan- dered into his former areas occasionally(1965 records), but resetfledwith a new mate to the south. Male BX shifted northeast with a new mate 1964 to 1971. In many other cases,new mates or new neighborscaused minor shifts in territories. Femalesshift gradually when under six monthsold, but eithersettle or shift drasticallywhen old. Young femalesare very evasiveand low on the peck order the first few monthsthey are independent.They are even lesslikely than youngmales to showconspicuous alarm behavioror agonisticdisplays. However,one youngfemale sang a few timeswhen about 75 daysout of the nest. The femaledoes not pair the first few monthsshe is inde.pendent, even if a territorial unmatedmale flirts with and feedsher assiduously.She may consortwith a male for severaldays, accepting his food, but she doesnot nest or singor attackintruding females. Oncesuch an intrudingfemale became the mate of the male, and the youngfemale wanderedaway. Some femalesnest as early as six monthsof age, as in BicoloredAntbirds, but I have no record 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 87 of a youngfemale Spotted Antbird breedingin the breedingseason in which she was hatched. Nearly all are mated and breedingin the next breeding seasonafter their hatching. After a female gainsa mate, she singsreadily to him or in confrontations with neighboringpairs, challenges and chasestrespassing females, and is quite noisyand adult in alarm and other behavior. She stayswith her mate in one area as longas he survives.When he is on the nestincubating, she sometimes wandersto feedat ant swarmsin nearbyareas; she is muchless limited to the territory than he is. If the male disappears,the female either getsa new mate and staysin more or lessthe samearea or desertsit completely,settling with an unmated male on his more or less distant area. Thus the generalarea usedby a pair is determinedby the bird that was therefirst. A newfemale or persistentlack of a femalecommonly causes minor shiftsbut rarely a major one, a new male causesno shift, persistentlack of a male causesa completeshift, and a settledpair staytogether and shift the area usedrelatively little over the years.

DOMINANCE AND TERRITORIALITY On his area the male is completelydominant; he challengesand chases trespassingmales, especiallyones that sing at or challengehim. He challengesand drives away trespassingfemales unless they chirp at him. Tres- passersthat returnpersistently and are quietor usethe cringingdisplay may be ignored if the male is foraging or otherwiseoccupied, or if the trespassers stayout of hisway, but directconfrontations start new attacksor challenging. As in Bicolored Antbirds, there is a reversal of dominanceas soon as the male leaveshis area and entersthe contiguousareas of other males. The male is now subordinate,even if he is older than the neighbor. The "age rule" of dominance,that older birdsdominate younger ones, has beenreplaced by the "territorial rule," that birds in their own areasdominate others. The dominancereversals with spacesuggest that the home areasof SpottedAnt- birdsare territories,by the definitionthat a territoryis "a spacein whichan animalor groupdominates others that dominateit elsewhere." The territorial boundaries,or zonesof reversal of dominancebetween territories,are rathersharp and stablein SpottedAntbirds. At timesmales chase each other back and forth acrosssuch boundaries;more commonly, males sing back and forth acrossthe boundaries.A male comesto playbackof tape-recordedsongs within his area and towardthe boundaryof his area,but singsto playbacksnear the boundary. Females challengeand chase trespassingfemales, and the latter chase them back when the combatants have crossed the territorial boundaries. How- ever, femalesare not as combativeas males,and they tend to be subordinate to trespassingmales. Of 1,194 recordedsupplantings and displacingswhere 88 ORNITHOLOGICAL MONOGRAPHS NO. 10 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 89 the sex of both SpottedAntbirds was recorded,805 recordswere between males, 213 betweenfemales, 150 were recordsof males dominatingfemales, and26 wererecords of femalesdominating males. In thelast cases, the female wasalmost always an adultfemale on territorysupplanting or displacingher own sonor a trespassingjuvenile or first-yearmale, and in mostcases her own matewas also present and chasing the trespassing male about. On oneoccasion the residentfemale snarledat an intrudingmale three times,then retreated as he quietly foragedher way. A residentfemale being crowdedout by an intrudingmale sometimessnarls faintly now and then, but is ignoredmuch of the time. One youngmale snakedhis neck out toward the snarlingresident female,and shestopped challenging. On severaloccasions when the resident male was not present,a wanderingmale supplantedthe residentfemale or chirpedto her as both occupiedthe swarmwithout disputing. In somecases, a returningyoung male supplantedhis own mother. Residentfemales some- timeswander into neighboringterritories, especially when following army ants, becausea neighboringmale is likely to ignorethem if they flirt with him and becausethe neighboringfemales are not very persistentat chasingthem.

TERRITORY SIZE, DENSITIES, AND BIOMASSES The pairs of SpottedAntbirds on the studyarea in Augustof 1964 are shownin Figure 20. In 1964 and other yearsthere has beenabout one pair of SpottedAntbirds every 4.7 hectares(12 acres), or 21.3 pairs per square kilometer. In May, just before the first youngof the year leave the nest, thereis onewandering bird per 50 hectaresor so,or two per squarekilometer. In May eachyear there are about45 residentand wanderingbirds per square kilometer, or some700 birds on the island of Barro Colorado. Sinceeach bird weighs15.4 to 19.0 g (exceptingtwo egg-layingfemales at 21.7 and 19.9 g), and the averageweights for 16 adult males, 10 other adult females,2 juvenilemales, and 10 juvenilefemales were 17.5 g, there are someeight g per hectareas comparedto three or less g per hectarefor the Bicolored Antbird (Willis, 1967).

HABITAT SELECTION On Barro Colorado, SpottedAntbirds are relatively common everywhere. However, they are most abundantin the youngforest and in certainregions

Figure 20. Pairs of Spotted Antbirds on territories on Barro Colorado Island in August, 1964. Names of pairs are placed approximately at the center of recorded observations(see Figure 19). Trails are marked off in 100-m sections,which are numbered on the map every 500 m. The laboratory of the SmithsonianInstitution is at the upper right. 90 ORNITHOLOGICAL MONOGRAPHS NO. 10 of old forest. The wide areasof old forestwith openundergrowth near the end of WheelerTrail, from Armour 8 to 20, and the canyonsin old forest betweenBalboa and Wheeler Trails do not seemto attract either wandering immaturesor territorial adults. Adults in suchareas have huge territories. Areasof old forestwith densepatches of Ananasmagdalenae, such as Zetek 9 to 17, have large populationsof SpottedAntbirds. So do vinecrowded youngwoodlands with occasionaltreefalls in the vicinity of Wheeler 5 to 15 and along Barbour Trail. The impressionI haveis that if one standsin the undergrowthand sees mostlylarge trunksand leavesof scatteredsaplings, he is likely to be in a placewhere few SpottedAntbirds wander or reside. If he can scarcelysee the bases of trunks for the leaves of low bushes or Ananas or for treefalls or vines or lianas, SpottedAntbirds are likely to be numerous. However, someareas of youngwoodland have rather open undergrowthand still have many SpottedAntbirds. In theseareas, I look for bird flocks and for treefall areaswhen I am lookingfor SpottedAntbirds.

DISCUSSION The lack of adultbehavior patterns, including alarm patterns, in wandering immature SpottedAntbirds probably has both morphologicaland environ- mentalbases. Exceptions to the generallack of alarm and agonisticreactions, suchas the youngmale (p. 84) that got a territoryand matedearly, suggest that immediateenvironmental pressures are more importantthan inability to performactions because of lack of structuresor practice. Whetherbecause of environmentor heredity,the lack of conspicuousalarm patternsin young is probablyenforced by natural selection.Young birds do not associatewith matesor youngthat couldbe warnedby alarmbehavior, so their offspringor potentialoffspring will not be helpedby callingor displaying.Moreover, young birds that gavealarm calls or displaysmight be unable to escapean alert predatoror might bring up territorial adults to competewith them or drive them away. Any advantagesthe youngmight gain by distractingor warningthe predatorare presumablyoverbalanced by the above and other disadvantages. Probablythe youngbird has little to gain and much to lose by singing or "challenging"at adults. If it keepsquiet, adultsmay overlookit. There is no reasonfor it to sing,for it hasno mateor young. It probablywould lose challengingduels with adults,which showvery aggressivebehavior if chal- lengedon their establishedterritories. The fact that youngmales tend to settlein a regioneven before they start singingand aggressivedisplays suggests that they gain by settling,perhaps becausethey familiarizethemselves with placesto hide from predatorsand 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 91 becausethey can estimatelocal competition.Probably the occasionalshifts to a new areasoccur when local competition reaches too high a level to start singingor challengingto set up territories;however, more detailedstudies are needed. Adult malestend to shift gradually,mainly when there is a changeof mates. Their statusis alreadyestablished in a givenregion, and to start out in completelynew regionswould probably reduce them almostto the low statusof wanderingyoung birds as well asexpose them to predationbecause of unfamiliaritywith the newareas. They are not muchmore likely to find a new mate more easilyby moving,since the settledfemales around them are generallyalready paired. The silentand mobileunmated females probably wouldcome to a male that stayedput and sangmuch more rapidlythan he couldfind themby wanderingquietly off his territory. Unmatedmales sing and wander more in their home areas than do mated males, thus increasing chancesof femalesfinding them or their findingfemales. It is moredifficult to understandwhy femalesstarting to form pair bonds at undersix monthsof age do not nestand singand drive awaytrespassing females. A female that finds a male and territory immediatelycan stay unless,of course,a more matureunmated female comesalong and displaces her. Perhapsthe youngfemale is not readyto fight or nest;her learning foraging,antipredator, and other abilitiesmight be slowedif she became occupiedprematurely with agonisticor reproductiveactivities. Since the femaleis rather unimportantin territorialactivities, only intrudingfemales needbe chased.Apparently the dangerof intrudingolder females must be slight,or the disadvantagesof beingdisplaced and havingto seek another mate are relativelyunimportant, since the young female does not react very vigorouslyto trespassingfemales. It is not evident why the young female doesnot attack,unless fighting would tend to interferewith other activitiesor directlyeliminate aggressive young females from the population. Perhapsmovement (sampling territories and availablemales?) is an advantage to the young female, however. At times adult malesshift accordingto the preferencesof the new mate, as in the case of male YBXR (Figure 19), and as in the Song Sparrow (Melospizamelodia) pair in whichthe maleenlarged his territoryto include a nesthis matebuilt outsideof it (Nice, 1943:191 ). Probablythe male sets the generalarea but the female can modify it, adaptingit for the special requirementsof a mated and nestingpair within limits set by neighboring pairs. Adult females,unlike adult males,leave the territory after losing the mate if no new mate appearswithin a short time. The female is subordinateto neighboringmales if she has no mate to protecttheir territorial dominance, hence will be subordinatewhether she stays or moves. By moving, she 92 ORNITHOLOGICAL MONOGRAPHS NO. 10 probablyexposes herself to predationbut definitelyincreases her chancesof findingan unmatedmale with a territory. That the presenceof a territorial male gives the female confidencein displacinga trespassingmale, and that flirtingsometimes stops male aggression,reminds one of socialpatterns in somemonkeys, where the offspring of a dominantfemale tend to be bolderwhen she is near, and "presenting" (asin baboons)is oftena gestureof a subordinate.Such patterns are probably likelyto arisein animalsthat are stronglydimorphic sexually or by age. The plaineror smallermorph is not equippedto outfacethe brighteror larger morph in competitiveagonistic displays, hence developsbehavior patterns that evade the disadvantage. There is currentlymuch discussionof the reasonsfor sexualdimorphism and associateddiethism (such as subservienceand flirting by femaleSpotted Antbirds) in animalsand humans. Verner and Willson (1966) noted the extremedevelopment of sexualdimorphism and the "doublestandard," in which males are polygamousbut femalesstay with one mate, in birds of l•abitatsthat have local placesof high productivity.Orians (1969b: 602) further suggestedthat, in animalsin which the male cannot (mammals) or neednot (locallyfood-rich environments) take careof young,that malesare likely to becomepolygamous, to fightwith othermales for accessto females or to localizedenvironments, and consequentlydevelop hypermasculine colors, size and behavior. Selander(1966: 141) suggestedthat on islands,with reducedinterspecific competition and many available niches, woodpeckers tendto develophigh sexual dimorphism to allowmales and femalesto spread into differentniches. Jehl (1970) suggestedthat for the Arctic, with its short breedingseasons, sandpipers have developed sexual differences in bill sizeto permitquick recognition of sexesand rapid pairing. It is interestingthat a commonfactor emphasizedby theseand other bi- ologists(Crook, 1964; Lack, 1968; etc.) hasbeen that irregularenvironments, with locally or temporarilysuperabundant foods, lead to sexualdimorphism and diethism. Severalstudents of humanshave also suggestedthat environ- mentswith irregulardistributions of resourceslead awayfrom egalitarianism both in economicsand in the rolesof men and women. Lewis (e.g., 1968) suggestedthat in humansthe uncertainand changeableslum subculturefavors care of offspringby femalesand hyperfeminityand hypermasculinity.Ap- parently the uncertaintyof jobs, except serviceones to dominant members of society,the fact that somemales have more money than othersand that some others can pretend importance,favors sequentialpolygamy, male strutting, and a complex of related factors much like the ones Verner has suggestedfor Long-billedMarsh Wrens. Glubb (e.g., 1963: 145) has suggestedthat the warlike but chivalrous,hypermasculine and hyperfeminine cultureof the Bedouinsof Arabia maybe adaptivefor their periodic,varying 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 93 sourcesof pasturage,as wasthe similarculture of chivalryin feudalEuropean societiescut off from tradeby theArab controlof the Mediterranean.Where life is uncertain,the masculinebut chivalrousfighter can win pasturageor his neighbor'sanimals or women. By this view, the presenttendency to sexual equalityand awayfrom heroicsand warfareis more adaptivein evenlypoor or evenly rich countriesthan in "developing"countries, because the last haverich areasnext to poor ones. Perhapsfemale or young animalsand humansin irregular or uncertain habitatswould be highlydisadvantaged at gettingfood if they did not have differentappearances or behaviorthan the males. Certainly,the tendencies to sexualdifferences in care of the offspring,in plumage,and in courtship seemgreatest among birds of variable or peripheralhabitats and least among birdsof matureor evenhabitats. The latterhave sparse but evenfood supplies splitup by competingspecies, and are perhaps most efficiently used by mated pairs workingfor food the sameway rather than by sexuallydimorphic birds workingin differentways. Spot-backedAntbirds, so closelyrelated to SpottedAntbirds that some ornithologistsmight place them in the same species,are less sexuallydi- morphicthan are SpottedAntbirds. The female Spot-backedAntbird has a white throat, but otherwiseboth sexeslook like slightly dull (somewhat female-like) but spot-backedversions of male SpottedAntbirds. The presenceof strongsexual dimorphism in only one of a pair of closely related speciessuggests that dimorphismis not genericallyfixed, but related to differencesin specific niches. Reproductivedifferences between Spot- backedand Spottedantbirds are little known,but are unlikelyto be striking. Skutch(1969: 292) recordedmutual grooming in Spot-backedAntbirds, and I have noted severalcalls and behavior patternssimilar to those of Spotted Antbirds. The main differencesI have noted are ecologicalones, primarily that Spot-backedAntbirds seldom follow army ants. It may be that irregular ant-followingby SpottedAntbirds is the aspectof theirenvironment that favors sexual dimorphismand diethism. If ant-followingrather than genericrelationship is importantin the question of sexualdivergence, one shouldcompare the moderatelyclosely related Bi- colored Antbirds and Spotted antbirds. Bicolored Antbird femalesbehave and look much like males. Indeed, male-male pairs sometimesform and persistfor years,so the birds themselvesclearly do not distinguishthe sexes well. BicoloredAntbirds are moderatelylarge (30 g), and are thusdominant enoughthat few speciescan drive them away from swarmsof army ants. They dependon ant swarmingto providea steadysource of food. Spotted Antbirdsare small (17.5 g) and are regularlychased away from the good central regionsof swarmsby BicoloredAntbirds and many other species. They cannotdepend on ant swarmingto providea steadysource of food. 94 ORNITHOLOGICAL MONOGRAPHS NO. 10

Instead,they forageat local or irregularly"available" sites (i.e., not pre- emptedby largerbirds) overthe antsor forageaway from ants(see below). How couldsexual dimorphism and diethismbe favoredby catchingcon- centratedbut irregularlyavailable food over antspart of the time? The most obviousresult of dimorphismand diethismis that SpottedAntbirds do not drivetrespassing birds of the oppositesex away from swarms,or splitup the swarm with them, as stronglyas do BicoloredAntbirds. Instead unmated male and femaleSpotted Antbirds can congregateat smallbut rich sites,or move quicklyto irregularlyavailable sites, if they happento be present together. A territorial speciesis thus transformedinto a semi-territorial species.When the local territorialmale is on the nest,his femalepermits wanderingyoung males or evenan occasionalneighboring male to trespass. Her lossin beingsubordinate to the trespasseris overcomein part by his susceptibilityto flirting and his tendencyto let her stay. The wandering youngfemale, the wanderingmated female whosemale is on the nest, and the femalethat has lost a mate are all toleratedby territorialmales to some extent. If the matesof theseterritorial males are present,there is little place for wanderingfemales; but even territorialfemales are rather "femininely" nonaggressivein SpottedAntbirds. Even when the trespasseris of the same sex as the residentbird, movementto a different corner of the swarm, off to foragenot far from ants,or waitingfor habituationto intrudersis sufficient to permitforaging in a speciesnot obligatelytied to ants. The habituationof aggressionin local males and femalescan be rapid if a different-lookingbird is involved,especially if that bird chirpsand flirts and thus arousesincompatible sexual tendencies. Spotted Antbirds need rapidhabituation, for theyhave such small territories and moveout of them so little that a wanderingbird followinga colonyof ants is likely to en- countera new pair of SpottedAntbirds each two or three days. Bicolored 'Antbirds,all of whichhave large territories and foragingranges, encounter new pairs much less often. The sexualdimorphism of SpottedAntbirds perhaps is a compromise,one that permitsfacultatively tolerant ant-followingwhile still maintainingthe ratherrigid territorialityneeded for useof evenlydistributed food awayfrom ants. Suchfacultative toleration of trespassersis probablydependent on the food supplybeing very irregularfor SpottedAntbirds at swarms. If food is likelyto changefrom one moment or siteto another,it maybe moreefficient for individualSpotted Antbirds to allowothers in closeto rich feedingsites, drive them away from intermediatesites, and avoid them at poor sites. Mobilityand flexibilityof responsemay be moreimportant than unvarying defense.The loweringof the generallevel of agonismby onehalf, by simply puttingmales and females into "differentspecies" in behavioraland morphological characters,may also attract less attentionfrom dominantbirds like 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 95

BicoloredAntbirds when the SpottedAntbirds do find a rich site. However, the level of agonisticbehavior is still high amongmale SpottedAntbirds, and fairly high among females. Probably any "altruism" toward trespassersoc- casionedby the sexualdimorphism is permittedgenetically because in many casesthe trespasserswill be relatives,and permittedecologically because the trespassersgenerally will be helping exploit locally superabundantfoods. Male SpottedAntbirds have developedsomewhat brighter colors and strongersongs than Spot-backedAntbirds, suggestingthat there has been an increasein male "assertiveness"in SpottedAntbirds as well as a decreasein female assertiveness.Possibly male SpottedAntbirds are better able to chase competitivemales away from their matesor goodfeeding sites at ant swarms if they have bright colors. There mustbe many disadvantagesto loud songs and brightcolors, such as increasedconspicuousness to predators, so that the assertivenessof males must be favored by their greater ability to drive away trespassersif competitiondoes become serious. To summarize,evolution of sexual dimorphismand diethismprobably is favored in irregular environments,such as alternatingbetween ant-following and ordinaryforaging, for many reasons. Reasonsfor one sex becomingless brightly colored, less assertive,include the following: antagonismis likely to be unusuallyhigh becauseof concentrationat rich local food supplies, and loweredassertiveness in one sex is a quick way of loweringantagonism and permittingtemporary use of the rich supply;lowering antagonism means that fewer competitorsof the same or other specieswill detect combatants and home on the rich food site; wanderingindividuals (often geneticrelatives of the territorial ones) are less likely to be driven off, thus permitting their survival under fluctuatingconditions that force them to wander, and allowing useof rich food sourceslike ant swarmsday after day; smallsize of "territories" comparedwith the movementof or geographicalchanges in resourcescompels a wandererto encountermany others,and different or nonaggressivebehavior or plumagefacilitates quick habituation;females can sneakin quickly at rich food sites;and femalescan be protectivelycolored for incubationor other purposes. Reasonsfor the other sex becomingmore assertivehave mostlybeen discussedby Orians (1969a) and others,and include: males (or females) can grab goodsites if they are brightly colored (etc.) and win in local fights with others; in an ephemeral or local environment, females can be stimulated more quickly and recognizemales of the right speciesmore easilyand quickly; and the largersex can sometimesprotect against the suddenlyappearing predators characteristicof an irregular environment. Factors acting both ways, to decreasethe assertivenessof one sex and increaseit for the other, includesome of the aboveplus the low interspecific competitionof irregularenvironments, which favors spreadof niche by sexual 96 ORNITHOLOGICAL MONOGRAPHS NO. 10 specializationor use of irregularlysuperabundant food in the way Spotted Antbirds use ant swarms. One must emphasize,however, that an ant swarm that is irregular to a SpottedAntbird is not at all irregularto a BicoloredAntbird; one often has to studya speciesclosely to know whetherit facesirregular conditions or not. Of course,the irregularconditions that favor strongsexual dimorphism and diethismare often much more apparentthan they are for SpottedAntbirds, which are not very dimorphicor diethic. SpottedAntbirds are stronglyterritorial and drive trespassersoff rapidly. They thus actually show aggressiveand submissivebehavior less often than do Bicolored Antbirds, which permit trespassersnearby but show much agonisticbehavior. SpottedAntbirds can forageaway from swarmsof army antsreasonably well, hencedo not have suchstrong reasons for stayingnear a domineeringresident pair at a swarmas do trespassingBicolored Antbirds. The trespassers'lack of persistencemakes it possiblefor the residentto chase them off with little wasteof time or energy. When a SpottedAntbird returns persistently,it is eventuallytolerated to someextent. Presumablyaggression or submissionare selectedfor only when they gain for an animal more than they lose for it. They can be selectedfor mainly in a moderately localized (irregular) environment, since in a widespread environmentor an extremelylocalized one efficiencyat gettingfood is better than efficiencyat defending,or staying,and territorial or other systems evolvethat minimizethe necessityof agonisticbehavior. I suspectterritoriality limits aggressionrather than being basicallyaggressive itself. The fact that amongterritorial birds submissionin neighboringareas lowers the expression of aggressionthere certainly lowers the generallevel of aggressionin the population. If there is goingto be aggression,it may be better to have territoriality as a brake on it rather than to have it expressedin fightsfor domi- nanceevery time two animalsmeet. If an animalthat cannotgain by aggression is able to developnonagressive interactions, of coursethe territorial system and its associatedlocally submissivebehavior can be dropped.

FORAGING BEHAVIOR The first reactionsof a SpottedAntbird when one approachesand tries to studyits foragingbehavior are likely to be chirring and mobbing,flight to cover, or chippingand panicking;less striking reactionsinclude singingas the bird movesover to the mate. Ordinarily the bird beginsto forage more andmore regularly, and soonalmost ignores the observer.Birds watchedover the years have becomefairly tame, so that they scarcelychirr even when I first appear. Data in this section,tinless otherwise specified, represent observationsof birds that were neither showingalarm patternsnor other non- foragingactivities, such as feudingor bathingor preening. 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 97

805ß 25'

2O ß Prey (n=396) \ /o o Perch before attack(n=106) + Perch (n=849)

-I-

ß ß ß \ /o

.... o15' ' i.o 2 3 4 g Height in Meters Figure 21. Heights of foraging away from swarms of ants: heights of perching, of perches taken before trying for prey, and of prey items tried for.

PERCH SELECTION

Foraging Spotted Antbirds generally wait on perches low in the forest undergrowth(Figure 21 ). They frequentlytake perchesnear the verticaland horizontal (Table 2). Perches over 4 cm in diameter are seldom taken, whether the perchesare under 45 degreesfrom the horizontalor not (Table 3). However,the birds sometimeshop alongsuch perches or on the ground for short distances. The small Spotted Antbird, with feet about 3 cm long, usesperches of about the same diametersas does the larger BicoloredAntbird, which has feet about 4 cm long. Apparently neither clingsvery effectivelyif the claws cannotgrasp half the circumferenceof the perch, but SpottedAntbirds cling to relativelylarge perchesmore often than do BicoloredAntbirds. The latter occasionallycling to large perchesby repeated graspingor sliding with the feet or by flutteringthe wings,especially when capturingan insect. Spotted Antbirds seldomdo so, perhapspartly becausethey are lighter and often can hover to get suchprey or neednot supportsuch great body weights. A relative of the SpottedAntbird, the Scale-backedAntbird of Amazonia, is a smallbird that specializesin clingingto perchesof small to large diameter for long periods. It has feet about 3.5 cm long. The SpottedAntbird seems 98 ORNITHOLOGICAL MONOGRAPHS NO. 10

TABLE 3 ESTIMATED DIAMETERS OF PERCHES OF SPOTTED ANTBIRDS

Foraging Away Preening from Ants Foraging with Ants Diameter, Per- Cm Records • Percent Records • Percent Records cent

1 16- 25 15.8-18.1 72-127 30.5-35.6 24 31.6 2 38- 61 37.6-44.2 93-159 39.4-44.5 30 39.5 3 31- 42 30.7-30.4 36- 54 15.3-15.1 16 21.1 4 11- 9 10.9- 6.6 15- 9 6.4- 2.5 3 3.9 5 1- 0 1.0- 0 15- 6 6.4- 1.7 2 2.6 6 2- 0 2.0- 0 7 0- 1 0- 0.7 8 2- 0 2.0- 0 2- 0 0.8-0 10 1- 0 0.4-0 10+ 2- 2 0.8-0.6 1 1.3

Total 101-138 236-357 76 Perchesunder 45 ø angle, followed by perchesover 45 ø angle. to be more of a generalist,intermediate to the Scale-backedAntbird and the BicoloredAntbird, in the diameterof perch relative to body size. The feet of thesethree birds are very similar (specimensexamined at the AmericanMuseum of Natural History). The toes and claws of Bicolored Antbirds are lessslender than those of Spottedand Scale-backedantbirds, a relationone mightexpect from the greaterbody size of the BicoloredAntbird. The solesof the feet all have roughenedor "cobblestone"surfaces, which perhapsgrip the percheswhen the birds clamp the feet but slide easilywhen they swingaround perches. Tarsal lengthsare about the samein all three species,varying from about22 mm in SpottedAntbirds to 24 mm in Scale- backedand 27 mm in BicoloredAntbirds. The differencesin clingingbetween Spotted and Scale-backedantbirds and the similaritiesbetween Spotted and Bicolored antbirds do not correlate well with the morphologicalseries of Spotted to Scale-backedto Bicolored antbird. The long and slender toes and short hind claw in the Scale-backedAntbird, relative to the size of these partsin the slighfiysmaller Spotted Antbird, may showmorphological adaptationsfor clingingwith the bodynearly level on verticalperches. SpottedAntbirds usuallyforage in moderatelyopen undergrowth,near a vine tangle or fallen tree into which they flee if the observer or passing hawks disturbthem. In very open undergrowththey move rapidly and are difficult to watch; they do not forage much. Dense growth, such as thickets of wild pineappleand the interior of treefalls,serve mainly as preeningor hidingplaces; the birdsseldom forage unless ants are flushinginsects. 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 99

TABLE 4 ATTEMPTS AT PREY BY SPOTTED ANTBIRDS AWAY FROM ANTS 1

Attempts at Given Height (Meters) Prey Location 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 2 3 ? Total

Air S2 I I 2 4 L a I 1 2 Leaf, twig S 1 4 2 3 2 4 1 1 11 4 4 37 L 1 211 2 5 12 Stem, liana, limb S 4 1 3 2 2 3 1 1 3 14 34 L 2 2 23 I 51 3 19 Trunk, log S I 4 5 L I 1 Ground S 290 290 L 9 9 T 4 10 10 Unspecified S 1 3 4 L 1 1 ? 1 1

Total 316 3 1! 3 9 4 6 8 2 3 27 6 31 429 • 1 October 1960 to 30 September1961. • Sallying (n = 374). 3 Lunging (n = 44). Leaf tossing(n = 10).

FORAGING AWAY FROM ANTS To forageaway from ants, a SpottedAntbird waits several seconds to several minuteson eachperch, looking around and down at the ground. For 28 intervalsbetween foraging attempts of birds foraginguninterruptedly, the averagewas 111.8 secondsand the extremes 8 and283 seconds.The foraging bird sits,half-sits, or stands. At timesit reverseson the perch or flies a few metersto a new perch. Passivewaiting is the rule, activehopping or clam- bering or flutteringthe exception.Now and then the bird lungesat prey, with or without a flight to a nearby location, or sallieslike a flycatcher to snapup the prey on the wing (Table 4). The most commonmethod, sallying to the ground,involves a quick jump or jumpingflight at prey on or near the ground,a brief flutteringcapture on or near the ground, and a quick bounceor bounce-flightback to a perch abovethe ground. The V-shapedattack and return takesless than a second unlessthe antbird hopsor spinsand fluttersto capturethe prey. If prey hidesunder the leaf litter, the bird sometimesstands on the ground with legssplayed and tail raisedslightly and tossesone fallen leaf after another. As is the casefor BicoloredAntbirds, each leaf is graspedin the bill and tossed 100 ORNITHOLOGICAL MONOGRAPHS NO. 10 one way or the other with a quick lateral motion of the head and neck or head and body. In salliesto vegetation,the SpottedAntbird is lighter and more agile than is the larger BicoloredAntbird. The former hoversreadily, if briefly, to snapprey off leaves,lianas, logs, trunks, and othervegetation or out of the air. Generally a SpottedAntbird is perchednear the ground before a try at prey,and most prey comes from the ground(Table 4). Only rarely,away from swarms,is prey taken as muchas two or threem abovethe ground. Once, however,a pair went to five m up to snapup wingedAzteca antsemerging in large numbersfrom a nest. Birds away from swarmsare not subjectto competitionnear the ground,or can easilymove to other equallygood sites if competitorssupplant them. Interspecificcompetition away from swarms of antsis discussedlater, underthe headingof "associationwith interspecific flocks."

SEARCHING FOR ARMY ANTS

Often one seesor hears singingor chippingSpotted Antbirds traveling through the forest undergrowth. At times the bird moves in one direction, perhapscommuting to a nestor goingto an antswarmit alreadyknows about. In the evening,after about 17:00, many of thesetraveling birds may have been goingto bathe or travelingtoward their roostingareas. At midday I less oftennoted travel. In the earlymorning, many birds were singingor traveling, or both,rather rapidly in irregularpatterns. They lookedabout and forageda little as they went, but they did not forage as efficiently as did settledbirds later in the day. I suspectthat a similarlyhigh rate of travelingamong Red- throatedAnt-Tanagers in the early morning(Willis, 1960a: 164-165) repre- sentssearching for swarmsof ants, and I suspectthat SpottedAntbirds also searchfor swarmsin the first few hoursafter dawn. If one bird of a wandering pair findsants, it stopssinging, then gives occasional songs that lead its singing mateto the ants. Birdsthat havenot foundants settle down to foragingaway from ants. On nine occasionswhen I played the recorded loud-songsof Bicolored Antbirds, Spotted Antbirds appeared. Probably Spotted Antbirds, which often associatewith Bicolored Antbirds at swarms of ants, use the calls of BicoloredAntbirds as cluesto the locationof swarms. In other playingsof recordedsongs, other ant-followingspecies came to the songsof Bicolored Antbirds and somecame to the songsof SpottedAntbirds (Willis, 1967: 25). In contrastto the looselyterritorial Bicolored Antbird, the stronglyterritorial SpottedAntbird seldom comes to the songsor recordedsongs of its own species unlessthe soundcomes from its own territory. Once a loud song-dispute betweentwo pairsof SpottedAntbirds brought up a wanderingfirst-year male, 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 101 but oneresident male quickly chased him away. More oftensong duels be- tweena pair at a swarmand a distantpair end in the departureof the distant pair, whichthen forages away from any swarm. As well as comingto the callsor songsof other ant-followingspecies, the SpottedAntbird reactslike any habitualant-follower to a "trail" (the ant highwaybetween bivouac or nest and the swarmraid) of Eciton burchelli. It followslow over the trail, perhapsgiving sharpchips as it goes,until it reaches the swarm of ants. When the bird chooses an ant trail that leads it to an inactiveraid, it reversesand follows anotherbranch trail to the main ant raid. SpottedAntbirds daily investigate a stataryant bivouacin their territory, andreturn periodically throughout the dayif no antsare swarming.If bickering flocks of BicoloredAntbirds are waiting near an inactive bivouac, Spotted Antbirdswander about nearby and investigate the antsperiodically. However, theystay out of the way of pugnaciouslarger birds, and they are not as prone to staynear the bivouacand rest as are birdsmore dependenton the ants. SpottedAntbirds are likely to wanderoff and startforaging elsewhere. When the ants do start raiding,these birds are likely to be an hour or two later at finding it than are more stronglydependent ant-followers. However, the SpottedAntbird returnsto a givenstatary bivouac in its territoryso persistently that it is likely to be the first or only bird presentif the antsdelay swarming until late in the day.

FOLLOWING THE ARMY ANTS The SpottedAntbird commonlyfollows army antsif any are availableand there are not too many dominantlarger speciesof birds about. Figure 22 indicatesthat there is little changein the percentageof birds at swarms during the year. A slightdecrease in the frequencyof attendingswarms at midyearprobably results in part from my flushingbirds away from undiscoverednests or young (and consequentlyaway from ants). Moreover, the wanderingimmatures common in othermonths often follow ant coloniesmore persistentlythan do territorialadults. In periodswhen I have searchedfor SpottedAntbirds away from swarms,such as duringcensus work from 1963 to 1971 (excludingJanuary-March 1965, when I was taking movies at swarms), I nearly alwaysfound a higher percentageof SpottedAntbirds away from swarmsthan duringthe first year of study. Censusesby someone workingwith SpottedAntbirds are needed,rather than by a personcom- mitted to studiesof birds at swarms. Probablymore than 50 percentof their food comesfrom swarms,and I doubtthat morethan 60 percentof foraging SpottedAntbirds are awayfrom swarmsover the courseof the year. The SpottedAntbird is the only bird I know that takes about50 percent of its food at swarms. Below it the next speciesis the Red-throatedAnt- 102 ORNITHOLOGICAL MONOGRAPHS NO. 10

o 1960 +196l 20 ß ß ' 1962 Y 1963 [] 1964 ß 1965 •, 1966

o Jan' Feb' Mar'Apr 'May 'June'July 'Aug ' Sept'Oct ' Nov' Dec' Figure 22. Percentage of foraging Spotted Antbirds recorded away from swarms of army ants and nests.

Tanagerat about20-30 percent(Willis, 1960a: 160). Above it the Barred Woodcreeperis at about 70 percent. I know very few birds in the middle rangeof the scale,between 10 and 80 percent,in contrastto the fair number of speciesabove 80 percentand the huge numberof casualvisitors below 10 percent. In the absenceof competitors,a SpottedAntbird foragesover ants much the same way that it forages away from ants. However, it often clings, pitches,and reversesactively rather than waiting patientlyas it does away from swarms. It moves from one vertical sapling or fallen liana or twig to another,to and fro along the swarm front. As a swarm movesahead, the attendingbird periodicallyflutters ahead, so that it keeps over the central section of the front much of the time. It is much like a Bicolored Antbird in its foragingbehavior over the ants. Of 1,692 recordsof foragingheight away from competitors,1,547 or 91.4 percentwere under one m and 1,091 or 64.5 percent were under 0.4 m (Figure 23). As is the caseaway from swarms,most birds are on perches near the vertical or horizontal,rarely on perchesoverhanging as much as 30 degrees(Table 2). Whether or not the perchesare below or above 45 degrees,most perchesselected are under three cm in diameter. Perchesof under one cm in diameterare used more frequentlythan they are away from swarms. Generally the perch is a bare one, with a clear view of the ground. SpottedAntbirds foraging actively at goodsites and awayfrom competing 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 103

IOO %

ocompetitors fl \ ,\ - ••X o ',,, .... •0 • •ompetitors present • •, n=1092 O•o•X • . I 2 3 4 5 6 7 Height in Meters Figure •3. •cJ•hts o[ [ora•iQ• Spotted A•tbJrds at swims o[ army aQts whcQ compeQn•lar•cr specieso[ arebirdswere presentaQd when they were abscm, larger birds tried for prey at averageintervals of 22.9 seconds(N = 41) over swarms of Labidus praedator and 43.2 seconds(N = 35) over swarms of Eciton burchelli,for a combinedaverage of 32.3 seconds.The interval for birds away from swarmsaveraged 111.8 seconds(N -- 28). The almost fourfold advantageof working over ants is less when competinglarger antbirds are present, however. As is the case away from swarms,most prey is capturedby sallyingto the ground(Tables 5, 6). At timesthe bird pecksat the leaf litter and quickly swingsback up, or standson the ground tossingdead leaves until prey is uncoveredand pecked. Only once did one dig deeply into the leaf litter, using its beak. Once one peered under a bark strip on the ground. Rather more frequentlythan away from swarms,and more often than doesthe larger BicoloredAntbird, the light and agile SpottedAntbird salliesinto the air or to foliageabove the groundto snapup a flying or fleeingarthropod. At other timesthe SpottedAntbird pecksprey out of the air or off foliage,either with or without a preliminaryflight to a point near the prey. However, unlesscom- petitorsare present,it seldomhops about peering and peckinglike an antwren or warbler.

PREY AND PREY TREATMENT At and away from swarms,most prey itemsare the length of the exposed beak (14.0 mm, or 1.0 B = 1 bill length) or less. Small items are swallowed 104 ORNITHOLOGICAL MONOGRAPHS NO. t0 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 105 106 ORNITHOLOGICAL MONOGRAPHS NO. 10

TABLE 7 PREY OF SPOTTED ANTBIRDS AWAY FROM ANTS 1

Estimated Size (mm) Kind ? 0-5 10 15 20 25 30 35 40 45 50 60 Unspecified 14 12 11 1 1 Sowbug 3 Spider 4 1 4 1 Arachnid 1 Centipede 1 Roach 1 1 Orthopteran 1 5 2 Moths 3 1 2 1 Caterpillar 2 1 3 1 2 1 Ant 2 1

Total 9 14 19 25 3 5 4 1 0 2 1 • Barro ColoradoIsland, 30 September1960 to 27 August 1969. sorapidly I wasseldom able to identifythem or evento tell whetherthe bird had capturedthe prey. Somewere sowbugs, moths, beetles, and spiders.Items juggledor chewedlong enoughthat I couldestimate their lengthsas fractions of beak length or identify them are listed in Tables 7 and 8. These items probably bracket the upper size limit of prey taken by Spotted Antbirds. Roaches,spiders, crickets, katydids, moths, caterpillars, centipedes, and others were frequentprey. One bird carried a lizard (/tnolis limiIrons) twice the lengthof its head, then droppedit. Prey over 1.5 B is seldomtaken, unlesslong and narrow like a caterpillar or centipede. Items over 1.0 B and some smaller items usually required specialprocessing. A SpottedAntbird never holds prey in the feet while peckingwith the bill. Commonlyit chewsmedium-sized prey or shakesit vigorouslybefore gulpingit more or less energetically.At times prey from 0.7 B to 3.0 B, especiallycaterpillars, are poundedor flailed on perchesas well as chewedback and forth. Moderateto large prey (0.5 B up) is some- timestaken to the ground,where the bird dissectsit in much the sameway as does a BicoloredAntbird. I rarely noted SpottedAntbirds dissectingprey away from swarms. At swarmsdissection was more common, but not so commonas in the large Bicoloredand OcellatedAntbirds.

NUMBERS AT SWARMS The numberof individualsat swarmsof ants changesrather little during the year (Figure 24). Young SpottedAntbirds enter the populationfrom 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 107

TABLE 8 PREY OF SPOTTED ANTBIRDS AT ANT SWARMS

Estimated Size (mm)

Kind ? 0-5 10 15 20 25 30 35 40 45 50

Unspecified 108(214) • 61(26) 53(18) 34(42) 1(3) 3(2) 1(2) Sowbug 5(2) 1 Spider 3(8) 6(2) 8(10) 1(3) (2) Spider egg sac ( 1 ) Scorpion 1 Centipede ( 1) 1 (1) Bristletail (1) Roach 3(1) 4(2) 12(21) 2(8) 2 2 Orthopteran (2) 3(2) 18(15) 4(3) 1(3) (1) Beetle ( 1) ( 1 ) ( 1) Caterpillar 1 ( 1 ) 2 ( 1) Moth 5 1 4 Ant 1(1) Lizard (Anolis) ( 1)

Total 125(229) 62(27) 67(24) 77(90) 8(19) 7(8) 6(4) 2(1) 1 1 Numbers outside parentheses, 1 October 1960 to 30 September 1961; numbers within, 1 October 1961 to 27 August 1969; Barro Colorado Island.

Juneto November. The numbersof swarmsof Labiduspraedator peak in the wettest months, October and November, and dip in the driest months, Januaryto April. Coloniesof Ecitonburchelli swarm all year. However,in March andApril somesplit into newcolonies. It is not knownwhen colonies of Labiduspraedator reproduce. Figure 25 indicatesthe numberof swarms per SpottedAntbird territory in 1960-1961; the estimatesassume that 40 percentof the swarmsof Labiduspraedator were missed during strip censuses (Willis, 1967; figure 2, p. 9) and that therewere 21.3 territoriesper square kilometer.The activitycycles of antsand birds should lead to an increasein numbersof SpottedAntbirds at swarmsfrom Juneto March, than a decrease until the first youngantbirds leave the nest the next June. There is seldommore than one pair of SpottedAntbirds at a givenswarm (Table 9). The residentpair normallydrives trespassing pairs away, or the latter take separatebranches of the swarm,so that thereis either temporal or spatialexclusion. Wanderingbirds are sometimestolerated at the same branchof the swarm,but generallyforage well awayfrom residentbirds and have to take the poorerforaging positions noted in the next section. Spotted Antbirds supplanteach other from long distances,up to 20 m at a time, sothat trespassingbirds of the samesex as residentsseldom forage near them. Intraspecificsupplantings are rather uncommonin Spotted Antbirds corn- 108 ORNITHOLOGICAL MONOGRAPHS NO. 10

o 1960 +1961

Jan Feb Mar Apr May JuneJuly Aug Sept Oct Nov Dec Figure 24. Spotted Antbirds at swarms of Eciton burchelli and Labidus praedator in 1960-1961.

.p.o '= ol960 •--0.8 '1961 TOTALo•ø,•, • ( Labiduspraedator •c• 0.6 œcitonplusburcheil/) o

= • •rchell/

O•n' F• ' M•r'•r ' M•y' Oun•'Ouly '•u• ' S•' Oc•' Nov' D•e ' Figure 25. Number of ant swa•s per Spotted Antbird territory in 1960-1961, as- suming21.3 pairs of SpottedAntbirds per squarek•ometer and using data on •t swa•s from Willis (1967:7-9). 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 109

TABLE 9 PAIRS OF SPOTTED ANTBIRDS AT SWARMS OF ARMY ANTS

Ant Species: Eciton burchelli Labidus praedator Pairs: Oq- I 2 3 4 Oq- I 2 3 1960 November 3 7 1 1 9 20 December 5 9 3 2 1 6 7 2 1961 January 8 12 2 6 February 12 14 7 4 1 March 8 9 1 2 2 3 April 14 17 6 3 3 1 May 22 20 4 1 7 2 1 June 22 13 4 4 July 28 28 2 5 5 August 18 28 2 2 6 1 September 20 25 5 5 4 October 21 28 6 6 10 1

Totals 181 210 30 6 1 58 74 7 paredwith BicoloredAntbirds (Willis, 1967: 43), becauseSpotted Antbirds drivetrespassers away so quickly that onesees few attacks.

COMPETITION AT SWARMS Many speciesof birds follow swarmsof ants,which are often narrow and crowdedwith individualbirds. At suchtimes interspecific and intraspecific attacksmay be frequent. Table 10 listsovert supplantings,casual displacings (when a bird obviouslyretreats from another), and returns (when a bird waitsuntil a largerbird leaves,then movesin to the areawhere it wasfeeding) of and by SpottedAntbirds on Barro Colorado. As in many tables and figuresin this report,the basicdata comefrom the first full year of relevant and relativelyunbiased observations (here 30 September1960 to 30 Sep- tember 1961); later observationsare perhapsbiased by being concentrated in certainmonths of the yearand in otherways, but are addedhere in separate columnsfor comparison.The tableincludes all of the recordedsupplantings, but onlydear and obviousdisplacings and returnsare listed. BicoloredAntbirds, twice the sizeof SpottedAntbirds, readily supplant or displacethem. A hissand a snap,and a supplantingBicolored Antbird takes the perch of a SpottedAntbird if the latter salliesfor food or waits 11o ORNITHOLOGICAL MONOGRAPHS NO. 10

TABLE 10 ATTACKS ON AND BY SPOTTED ANTBIRDS AT SWAP.MS OF ANTS

Species Supplantings Displacings Returns 1. SpottedAntbird 422(746) • 80(103) 1(o) 2. Bicolored Antbird 223(205) 183(80) 41(43) 3. Ocellated Antbird 30(37) 55(30) 13(11) 4. 2 and 3 3(4) 5(1) 26(20) 5. Gray-headedTanager 7 (3) 24 (3) 3(2) 6. Plain-brownWoodcreeper 5 (2) 5 (4) 7. Chestnut-backedAntbird 0(40) 1 (5) o(1) 8. Slaty Antshrike 2(2) 1 (2) 9. White-whiskered Puffbird 5 10. SquirrelCuckoo, 1 0(3) 11. Great Rufous Motmot 1 (1) 12. Scaly-throatedLeafscraper 1 13. Buff-throated Woodcreeper 1 14. Barred Woodcreeper 1 15. Swainson'sThrush 1 1/12 16. Gray-cheekedThrush 1 1/1 17. Veery 0/2 18. Thrush (species) 0/1 1 19. KentuckyWarbler 0/2(4) 0/1 20. Canada Warbler 0/0(2) 0/0(3) 21. Acadian Flycatcher 0/1 22. Ruddy-tailedFlycatcher 0/0(1) 23. Checker-throated Antwren 0/1 (1) 24. Red-cappedManakin 0/1 25. White-flanked Antwren 0/0(1 ) •30 September1960 to 30 September1961, no parentheses;1 October 1961 to 27 August 1969, in parentheses.: Recordsbelow slashlines indicatespecies subordinate. nearby.The two foragein muchthe sameway, and a BicoloredAntbird rarelypermits a SpottedAntbird withintwo metersof its perch. However, Table 11 listsa few recordsof BicoloredAntbirds permitting Spotted Antbirds to forageclose to them ("ignores")and of SpottedAntbirds foraging past Bicolored Antbirds ("infiltrates"). Generally the Bicolored Antbird was preeningor feudingwith anotherof its own speciesat suchtimes. At times, however,the BicoloredAntbird was foragingactively but still permittedthe SpottedAntbird to work nearby. Normally the larger speciesgrunts frequently,a call noteoften directedat any nearbysupplantable competitor, and supplantsthe SpottedAntbird aftera few seconds.If the SpottedAntbird tries 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 111 112 ORNITHOLOGICAL MONOGRAPHS NO. 10

TOP VIEW OF SWARM

Figure 26. Foraging zones around a swarm of ants, as seen from the top and front and side. "Zone a" is taken by dominant birds, "Zone b" by subdominants, and "Zone c" by subordinates. for prey whenthe BicoloredAntbird is not busy,the supplantingcomes more rapidly---evenas the SpottedAntbird reachesthe groundon a split-second sally. The even larger Ocellated Antbirds do not often supplant or displace SpottedAntbirds, becausethe smaller birds stay well away from the former muchof the time. Normally the dominantOcellated Antbirds take the center of a swarm,the subdominantBicolored Antbirds take a concentricbut shifting ring around them, and the subordinateSpotted Antbirds take a peripheral and shifting ring next to the Bicolored Antbirds rather than move near OcellatedAntbirds (Figure 26 "Zonesa, b, and c"). Many other large or moderatelylarge ant-followingspecies, such as the Gray-headedTanager, also supplantthe SpottedAntbird. In general, the effectof multiplesupplantings by Chestnut-backedAntbirds on the infrequent occasions when the latter follow ants is about the same as the effect of BicoloredAntbirds--the SpottedAntbird has to forageperipherally or at the other end of the swarm. Slaty Antshrikesfour times supplantedand three times displaced Spotted Antbirds, but the former. normally forage well above the latter at swarms. The Plain-brown Woodcreeperand the Barred Woodcreeperseldom supplant the SpottedAntbird, for woodcreepersgenerally foragehigh in the undergrowthor on pole-sizedor smallertree trunks in relativelyopen undergrowthand seldomwork the swarmat the sameplaces 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 113 or times as doesthe SpottedAntbird. Often the woodcreepersand antbirds ignoreeach other even when foraging close together. Occasionallylarge White-whiskeredPuffbirds, Great Rufous Motmots, or SquirrelCuckoos at swarmsflush a SpottedAntbird more or lessaccidentally by flyingdown past it in a chaseof somelarge prey. In casesnot listedin Table 10, passingcoatimundis three times displaced Spotted Antbirds, and a foragingScaly-throated Leafscraper ignored the antsbut supplanteda Spotted Antbird that was waiting over them. The SpottedAntbird occasionallysupplants passing smaller birds, such as the Ruddy-tailedFlycatcher and the Checker-throatedAntwren. These and other birds of the "antwrenalliances," wandering interspecific flocks, often drift near a bird flock at a swarmof ants,peer andflutter aboutin their usual fashion,or stare at the ants as if curiousand drift on. They are normally presentso briefly that they scarcelyare exposedto supplantingsby the larger ant-followingbirds. It may be that the antwrensavoid closecontact with large antbirdsbecause of occasionalsupplantings, but usually birds of the antwrenalliances seem to ignorethe antsand ant-followingbirds rather than avoidthem. Ordinarilymost of the antwrensforage higher in the undergrowth,2 to 10 metersup, than do the ant-followingbirds. The Canada Warbler is one small bird that readily shiftsback and forth betweenant-following and antwren aggregations. The SpottedAntbird readily supplantedor displacedthe Canada Warbler on the few occasionswhen a warblerforaged low enoughin the foliageto comenear one at an antswarm. Migrant thrushes,which rarely staywith antwrenflocks but readilyjoin ant- followingflocks (Willis, 1966a: 198), occasionallysupplant or displacethe smallerSpotted Antbird. However,Spotted Antbirds occasionally displaced or supplantedthrushes nearly twice their size. Whenlarge competitors exclude it from the bestforaging zones over the ants, a SpottedAntbird movesperipherally. Table 11 listsinstances of Spotted Antbirdsforaging in theseplaces: above,ahead of, and behindswarms or near the observer;behind logs or in densevegetation; infiltrating the ranks of larger antbirdsbriefly, or foragingpeacefully near them; waiting at the endsof swarmsor at separate(and usuallysmaller and rather poor) forks of swarms;and desertingthe swarmscompletely as direct resultsof the activitiesof larger birds. These data are taken from commentsin field notes and do not representorganized studies. Only a smallfraction of the observed caseswere recorded; the recordsare useful mainly as indicationsthat Spotted Antbirdsreact to largercompetitors in severalways and to showthe bases for Figure26. One problemis that the denseforest undergrowth makes it difficult to see birds ahead of the swarm and birds at the ends of the swarm as comparedwith birds behindthe swarmor near the observer. However, birds 114 ORNITHOLOGICAL MONOGRAPHS NO. 10 behind the swarmor near the observergenerally must be tame individuals. Birds at a separatefork or birds desertinga swarmare unlikely to be noticed more than onceduring a periodof observation,while birds foragingnear the observercan be recorded repeatedly. For Table 11, I recordedan individualas doinga givenactivity when that activity was separatedin my notes from the previous record of the same individualat the sameactivity by a record of that individualat someother activityor by observationsof someother bird. A bird that foragednear me for 20 minutes is counted once if I did not note that it moved off for a time or if therewas not a seriesof observationson other birds followedby a note that the given individualwas still near me. As categories,"wandering" and "deserting"refer to birds changingtheir locationsso much that they did not stay long in any one of the "location"categories; however, in somecases a bird wanderingbefore or after stoppingto forage,such as behindthe swarm, was countedas a record for wanderingand a record for behind the swarm. In view of the fair difficulty of seeingbirds ahead of the swarm, the total for SpottedAntbirds ahead of the antsand competingantbirds by one or more m is stillrather high. SpottedAntbirds commonly range back and forth ahead of the ants when interspecificcompetitors are present. The larger com- petitorsgenerally face the rear of the swarmrather than ahead, so that the occasionalsmall prey leaping ahead of the ants is available for Spotted Antbirds without great risk of supplantings.However, the area of good foragingis rather narrow ahead and to the sidesof the ants (Figure 26), becausemost prey itemsquickly hide undercover unless they are reflushedby the ants. "Foraging at the ends of the antswarm" generally means birds taking positionsdirectly over the ants,in caseswhere therewere not enoughlarge antbirdsfully to occupythe whole length of the swarm. In suchcases the large antbirdssometimes concentrate at the end of the swarmaway from the observer(Figure 26) and leave a large but very goodforaging zone at the end near the observer.At other timesthe large antbirdsmove back and forth or stay at the center,while the small SpottedAntbird takes whicheverend is momentarily unoccupied. There is a very wide foragingarea behindthe swarm,where the coalescing trails of ants form a wide "fan" with many foci of moderateactivity. How- ever, muchof this regionis rather poor for prey, sincethe antsand the birds aheadhave alreadycaptured much of the prey that has not reacheda safe hiding place. However, SpottedAntbirds do rather well searchingfor the last smallprey on the undersides of leaves,or work foci of activitynear the observerunless larger birds are tame enoughto forage there instead. Four timesSpotted Antbirds foraging behind the antsturned to "thievery," 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 115

or to stealingpieces of arthropodsfrom the homeward-boundarmy ants. In BicoloredAntbirds, thievery is restrictedto subordinateindividuals, ones that are gettinglittle prey becausedominant birds exclude them from good foragingareas (Willis, 1967: 30). In the four casesof thieveryI observed amongSpotted Antbirds, several Bicolored and other antbirds were occupying the mainswarm. In eachcase, the SpottedAntbird flew downto the ant trail, pickedup an antand booty, shook the ant off, andate the booty. None of the birdstried thievery more than once, although three watched the ant trail for a minute before leaving. Foraginghigh above the antsrequires hopping and peering, as the Spotted Antbirdinvestigates tangles or suspendedfallen limbs and otherprotected places.The Ocellatedand BicoloredAntbirds sometimes grunt at the small bird abovethem, and supplantit at timeseven when it is two or threem abovetheir heads, so foraging above other species is notalways safe. When two or morespecies of competingbirds are present,Spotted Antbirds forage highabove the antsless frequently than when only one speciesis present (Table11 ). Whenonly one competitor is present,it usuallytakes "Zone a" (Figure26) andleaves the other zones free, unless there are manyindividuals of its own species. If severalcompeting species were present,Spotted Antbirds were slightly less likely to forage near me; often tame subordinateBicolored Antbirds (generallythe secondspecies present) took this position(Willis, 1967: 62) at such times. A specieslike the Bicolored Antbird, which tolerates tres- passingindividuals until as manyas 15 attendedone swarm,often spreads out to occupyboth "Zone b" and "Zone c" or takesmuch of all three zones, forcingSpotted Antbirds to go to minorloci of activityat the outskirtsor to separatebranches or to desertthe swarm. Althoughthe SpottedAntbird can still find somefood by foragingabout the peripheryor over a swarmof antsif manylarger antbirds are present,it often desertsswarms of antsif there are many largecompetitors. One often findsSpotted Antbirds wandering around the bivouacor alongthe trail of ants rather than at the main swarm,or seesSpotted Antbirds only a few minutesduring s6veral hours of watchingat a swarm.Since the largeantbirds concentratearound the large and consistentswarms of Eciton burchelli,the SpottedAntbird is oftenrelegated to smallbranch swarms, to rather inactive raids,to the temporallyunpredictable statary raids of Ecitonburchelli, or to raids of Labidus praedator. If undergrowthis veryopen, the wary SpottedAntbirds often stay on the peripheryof a swarmor in densecover nearby, moving in over the antsonly occasionally.Their behavioris effectivelyalmost the same as when competitors exclude them. 116 ORNITHOLOGICAL MONOGRAPHS NO. I0

DISCUSSION PerhapsSpotted Antbirds forage relativelyseldom from perchesof low curvature(i.e., large radius), suchas fallen logsor the ground,because large perchesblock visibility,especially when thesebirds get most of their food from the leaf litter. Limited visibility might also block their view of ap- proachingpredators. Certainly, Spotted Antbirds usually forage in moderately openundergrowth rather than in tanglesthat couldblock their view of both predatorsand prey. However,they avoidvery openundergrowth, and often staynear tangles.The dangersof predationprobably rise as habitatsbecome very open or very dense. Perhapsperch sitesand prey are lesscommon in veryopen sites, while prey hide too easilyin verydense sites. Also, Chestnut- backedAntbirds occupy dense sites and chaseout SpottedAntbirds (Willis and Oniki, MS). There are manyspecies of antbirdsthat forageby hoppingor walkingon logs or lianas or the ground,including in the Panamanianlowlands such birdsas Streak-chestedAntpittas, Black-faced Antthrushes, and White-bellied Antbirds. The morphologyof the foot in suchclinging birds as Spottedand Bicoloredantbirds may hinderuse of suchhorizontal perches of low curvature. Neitherthe SpottedAntbird nor any of its relativesis very goodat hoppingor walkingprogressively, but their large feet and legspermit them to cling to verticalperches that hoppingand walking antbirdsscarcely use. Probably the long legsof the other antbirdspermit them to hop effectivelyor to stand up abovethe groundso they can seeprey or predators.The morphological adaptationsfor clingingand the behavioralavoidance of denseor very open sitesprobably restrict foraging opportunities for SpottedAntbirds; but predatorsand competitorsprobably would causelosses that would offset gains if SpottedAntbirds were lessspecialized. Swarmsare sometimesunattended even when there are no competitorsand nearby SpottedAntbirds are foragingaway from ants. Sincebirds foraging at swarmsfind prey four timesas often as birds away from swarms,I suspect such antbirds fail to find the ants rather than find them and leave them. It is possible,however, that someSpotted Antbirds desert the ants after getting sufficientfood or becausethere are no other birds about to help keep the lookout for predators. SpottedAntbirds searchfor ants mainly in the early morning. Probably they restrict their searchespartly becauseof danger from predatorsand partly becausethey cannotforage well while searching, sincethey forage best with relativelylong waits at perches. Searchingper- sistentlyuntil a swarm is located might be less efficient than foraging away from ants if it takes too long to find the ants; losseswould outweighgains. Bicoloredand Ocellatedantbirds, which are dominantat the easilylocated and regularnomadic swarms of Eciton burche'lli,search until they locatethe 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 117 ants;they find little of their food away from ants. SpottedAntbirds, low on the interspecificpeck order, cannotdepend on suchswarms. Where the ants areeasily located and competition is low,as at "statary"(Schneirla, 1957: 262) coloniesof Eciton burchelli,Spotted Antbirds do revisitrepeatedly during the day. Possiblyfood availabilityis more importantthan dangerfrom predators. The standardargument against predators' restricting foraging ranges of prey hasbeen the onethat a prey specieswill evolvebehavioral or othermechanisms that enable it to use the food in the presenceof the predator. However, in evolutiona speciesdevelops the abilityto exploitcertain foods and accordingly becomesabundant; then a predatorbegins to exploitit and reducesits numbers in peripheralor unsafezones. Then the speciesis replacedor evolvesfurther, and a new predatorcomes in or the old one evolves,and so on ad in[initurn. Alternation and lag effects are thus characteristicof predator-prey interactions, not evolution to the point where a speciesuses all possiblefood unaffectedby predators.This is probablyespecially true of complexsystems with numerouspredators, and less true of simple systems. In a complex system,the prey is restrictedin foragingbecause it must wastetime watching for a variety of predatorsor keep to safe sites,unless it is able to beat the systemby developinga high reproductiverate outsidethe system(an outlet available to migratory birds, which can be tame and exploit dangerous niches;Willis, 1966a: 222). The predatorstl•us limit the food supplyopen to any one speciesof prey, openingniches for other prey speciesand en- couragingspecies diversity indirectly as well as removingprey individualsfrom the populationdirectly (Paine, 1966: 65). Predatorsthus may limit the placessafe for a speciesand the kinds of food availableto it, much as do competitors,prey protectivedevices, environmental productivity,environmental complexity, different or varyingweather, and other factors. Speciesdiversity should increaseat first as any of these factors increase,but too much of each will then decreasediversity because specieswill be lost fasterthan they can be evolvedor movedin. It may be that greaterenvironmental stability, structural heterogeneity, or productivity doesnot give greaterspecies diversity ad in[initum but only up to a point. In a very productiveor perfectlyconstant environment, as well as in poor or variable environments,species will be eliminated. With high and constant productivity,for instance,specialized species can eliminategeneralists; with manypredators, some prey specieswill be eliminated.Margalef (1968:41 ) notesthat eutrophicationand high productivity usually leads to loweredspecies diversity;of course,very low productivityhas the sameeffect. The biological interactionsof predatorsand competitorsbecome more important than the physicalfactors when the latter are optimal; however,it may be that for maximum speciesdiversity there are optimal levels of food supply and of predatorsas well as of temperatureor rainfall, and that decreaseor increase 118 ORNITHOLOGICAL MONOGRAPHS NO. 10 in the first two awayfrom optimacan lower speciesdiversity just as can decreaseor increasein the last two. Aristotle's"golden mean" may apply to communitiesand speciesas well as to individuals. Perhapsthe SpottedAntbird's greater movement at antswarms,as compared with its activity away from ants, is causedby the fluctuatingactivity of the antsthemselves; the locationswhere prey is likely to appearchange rapidly as the antsmove, whereas one place may be aboutas likely as anotheraway from ants. However,this may not be quitetrue away from ants,as a SpottedAntbird movesperiodically; perhaps its activityeliminates or inhibitsthe likely prey at anyone place after a few minutes,or it desertsa placethat doesnot produce prey. It doesnot needto moveotherwise, except to keepup with thewandering antwrenflocks it often joins away from ants. The relativelack of prey dissectionin SpottedAntbirds as comparedwith their larger relativessuggests that the small arthropodstaken by the former are relativelyeasy to break apart. In all theseantbirds the averagemaximum lengthof prey correlateswell with the averagebeak length,so that there must be an allometricrather than a direct increasein toughnessof prey with length of prey. The great increasein size of bill among birds eating large insects (puffbirds, motmots) and very large seeds (Galapagosfinches, Bowman, 1961) probably compensatesfor an allometric increasein hardnessthat exceedsthe increasein prey length; but ant-followingantbirds seem to com- pensatemore by behavioralmeans, namely an increasein dissectiontime. Such increaseswould clearly be a disadvantageunder conditionsof strong competition,and shouldlead to the developmentof large beaks. This suggests that, even thoughthese antbirds compete for spaceat swarmsof ants, there is either not enoughcompetition to causeevolution of very differentbeak sizes or that the large antbirds can gain more food with small beak sizes even thoughthey have to spendtime dissecting.Since large organismsare much rarer than small ones in the leaf litter (Williams, 1941), there should be strong selectionagainst large birds or large beaks. Dominance,however, goesto the largestand (with increasein food per captureif prey are large) shouldprovide counterselection.Superior competitive abilities of the smaller birdsmay force the largerbirds to take largerprey to someextent. However, it is likely that the large birds take prey as small as they can efficientlyuse, consideringthat small itemsadd to capturetime more than to pursuittime (see MacArthur and Pianka, 1966: 603). The exclusionof smaller species by larger onescan be an advantageto the larger oneseven when they take different size rangesof food, as long as the larger ones gain more than they loseby chasing,and the dominance"zones" of Figure 26 are an extensionof morphologicaland behavioraldifferences between the speciesover swarms of ants. When no competitorsare present, Spotted Antbirds forage low over the 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 119 ants in much the sameways as do Bicoloredand Ocellatedantbirds. This suggeststhat this is the bestforaging zone for SpottedAntbirds as well as for larger birds. When larger antbirds exclude SpottedAntbirds, they forage very actively around the periphery of swarms. Although the great and temporallyextended activity of excludedSpotted Antbirds showsthat they are "adaptable"and relativelyunspecialized, with refugiato which they can turn, comparedwith nonexcludedbirds, their activity wastesenergy and time and exposesthem to predators. Moreover,each individualbird foragesover a wider zone,so that drivingaway trespassersbecomes more necessaryand populationsmust be lower. In extremecases, all the SpottedAntbirds forage away from ants. Suchsubordinate species probably must be relativelyunspecialized in any competitivesituations, or be locallyexcluded as what Hutchinson (1951: 575) hastermed "fugitive" species. At swarmsof ants,Spotted Antbirds are in a sensefugitive species, but they are rathersuccessful at findingnearby refugia andhave high population levels. The successof a speciesneed not be related to its degreeof specializationor to its dominancestatus, even if a constant environmentshould lead to eliminationof fugitiveand unspecializedspecies. Thereare plentyof irregularor locallyrich environmentsthat permitfugitive speciesand generaliststo be successful,even in the tropics. Competitiveexclusion by dominanceat swarmsof antssuggests that the variousspecies behave as competitorswhether or not theytake the samefoods. The large Ocellatedand Bicoloredantbirds probably use little of the small prey that the SpottedAntbird wouldtake, but they excludeSpotted Antbirds from their foragingareas as completelyas if the latter were anothermember of their own species.Intraspecific disputes and supplantingsare much more frequent and vigorousin all three species,but simple avoidanceof large speciesby smallones (Table 11) probablyaccounts for muchof the decrease interspecifically.Often the two larger antbirds start grunting, their notes for a supplantablecompetitor, and the SpottedAntbird fleesor movesaway. Thereis no needto fightor supplantunless there is resistance.Intraspecifically, matesand young are toleratedas closeas are membersof anotherspecies. Moreoverin OcellatedAntbirds independent young and their matesare also tolerated;and in both Bicolored and Ocellated antbirdspairs eventually tolerateother pairs within a few meters,or almostas closeas they tolerate youngor mates, as long as the subordinatebirds are submissive. It mustbe that all thisfighting for spaceat swarmsis possiblebecause there are actually local food surpluses,and hence subordinatespecies and subordinateindividuals of dominantspecies can stayand survive. If therewere a shortageof food, the best-fittedbirds shouldeliminate the othersor push them into other niches,and actualfighting or other evidenceof competition would be rare or anecdotalin nature (competitiveexclusion principle; 120 ORNITHOLOGICAL MONOGRAPHS NO. 10

Grinnell,according to Miller, 1964: 236). Withoutwatching the evolutionary or experimentalcourse of a competitiveinteraction, one is uncertainthat the disputingleads to separationof nichesand the statisticsare goingto be anec- dotaland can always be criticized.However, the fighting does seem to leadto a spatialpattern of separationof niches,even if fightingmay actuallyhinder evolutionof differencesin beak size (see below) or morphologicallybased niche differences. It is possiblethat disputing,even thoughit forcessubordinate species to evolveor be excluded,actually inhibits morphological evolution of a dominant species.If it canchase away the others,interspecific competition is eliminated as a selectivepressure for evolution. As other speciesrapidly becomemore efficient,the dominantone only turns to more and more aggressivebehavior. It becomesprogressively less efficient, and is gradually replacedby species that neednot wasteforaging or other (Ripley, 1959:132 "aggressiveneglect" of nests) time on aggression.More likely, judging from the birds at ant swarms,the other speciesmove in and the dominantones can no longer gain by attemptingto displacethem. The dominantsthen have to undergorapid evolutionto becomemore efficient,which may be an evolutionarycrisis and lead to local extinction. In either case, evolutionin optimal environmentsis probably toward eliminationof crude aggressionand toward institutionof morphologicaland other differencesnot related to aggression.This may also apply toward predatorsas well as to competition.Thorny and poisonousplants are less common in mature forest than in successionalor difficult habitats; human use of pesticidesseems self-defeating. In other words, the meek inherit the evolvedearth becausethe strongcan efficientlydefend only in an irregular world, such as over army ants at the presenttime, and then only before evolutionproduces meeker but quickerspecies. In an evolutionarysense, aggressionis eventuallygood for the downtroddenbut bad for the aggressor unlessthe downtroddenare quicklyand completelyexterminated. Since the competitiveexclusion principle statesthat monopoliestend to develop,one wonderswhy there are actuallyfew monopoliesin nature. One canhardly think of a typeof foodthat is eatenby one and only one organism, for instance.If one acceptsthe principlea priori, he mustlook for opposing principlesthat frustrateits completefulfillment. One obviouscounterprinciple is the "specializationprinciple," that morphologyor behaviorrestricts the range of thingsan organismcan use; it would be difficult for an antbird to monopolizeboth insectsand diatoms. The failure of Bicolored and Ocellated antbirds completelyto exclude Spotted Antbirds from swarms of ants is partly explainedby this principle, but as important is what I have called (Willis, 1'966b: 671 ) the "irregularityprinciple," that irregularitiesin time and spaceleave surplusesthat can be exploitedby quickly-movingspecies 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 121 that have alternatefood sources. When times are bad, or rather irregularly bad, they prosper. SpottedAntbirds should prosperin moderatelygood yearsfor food and in moderatelyirregular climates, more than in very good yearsfor food or in constantclimates. In very good yearsor in constant climatesbetter-adapted competing species should exclude them. Spotted Antbirdsare uncommonin the constantlywet partsof the Choc6of Colombia. They reachhigh abundancesin regionswith moderatelystrong dry seasons. However, they disappearin regionswith extremelystrong dry seasons. The Plain-brownWoodcreeper seems to occupythe samesort of peripheral and subordinatestations at the swarms,but has different refugia--high in the trees,areas of large-trunkedtrees and little groundcover, etc. (Willis, 1966b: 668). By the irregularityprinciple, it shouldbe lesscommon in areas of constantclimate than in areasof greaterirregularity, such as Barro Colorado. In constantclimates, specialists should outcompeteit. The speciesis uncommonin the Choc6 and in the constantlywet upper reachesof Amazonia comparedwith regionswith moderatedry seasons. There are, however, alternative explanationsfor these distributional patterns. Possiblythe SpottedAntbird and the Plain-brown Woodcreeper cannotnest or get enoughfood in suchregions for reasonsof high rainfall, or for variousreasons based on specializationsrather than on irregularity. The questionof whetherthe difficultiescome from the biologicalor the physical environmentare not yet answered.It is perhapsdesirable to seewhat ranges of conditionsa speciesoccupies and what changestake place when specified conditionsare changedrather than to speculateon the relative importance of competitionversus other factors. I have elsewhereproposed (Willis, 1966a: 217) that migrant birds are unassertiveand rarely competedirectly with residentbirds, even small ones like SpottedAntbirds, at swarmsof ants. A few attacksby SpottedAntbirds on thrushesof somewhatunder twice the weightare aboutthe only exceptions to the generalrule that larger birds supplantsmaller ones in interactionsat swarms.However, the recentlyanalyzed records of SpottedAntbirds foraging at the peripheryof a swarm rather than over it when many thrusheswere present (Table 11 ) indicate that there may be more effect of migrantson SpottedAntbirds than I thoughtin the earlierpaper, where I tabulatedonly obvioussupplantings and displacings.Moreover, it is possiblethat there were few young SpottedAntbirds leaving the nest in late 1960, a year of high numbersof thrushes,than in 1961. I saw only one young male Spotted Antbird in the fall of 1960, and many the next fall. However, in 1960 I was neither looking for young birds nor experiencedat detectingthem. There were many first-year SpottedAntbirds about from January to June, 1961. Still, it is possiblethat the unusuallylarge numberof swarmsof Labidus praedator in the wet year of 1960 were so completelyoveroccupied by 122 ORNITHOLOGICAL MONOGRAPHS NO. 10 thrushesand other birds that feweryoung Spotted Antbirds survived that year thanin the dry yearof 1961,which was a poorfall for antswarmsbut an even poorerfall for thrushes.There is thusthe possibilitythat SpottedAntbirds haveless success in nestingin yearsof abundantswarms, because of over- crowdingby migrants,than in pooreryears for ants. However, the number of such habitual ant-followers as Ocellated and Bicoloredantbirds and of variousresident occasional ant-followers may also have been high in 1960 and contributedto any lower successfor Spotted Antbirds. The availabledata (Willis, 1966a, figure 2) do not supportthis hypothesis;but in 1960 I wasinexperienced and probably scared away many residentbirds, especially in October. More detailedanalyses of the relativesuccess of nestingin SpottedAnt- birds and abundancesof migrantsand residentsin different periodsof fall migrationmay help determinewhether this speciesis actuallyaffected by the migrantsmore than I thought in 1966, when I assertedthat migrantsuse surplusfoods and generallydo not crowd residentspecies or affect their breedingunless they displacethem by nonaggressiveinfiltration. The Spotted Antbird is the ant-followingspecies that is most likely to be influencedby migrants,as almost all other ant-followingresident species are larger or more aggressivethan are ant-followingmigrants. Johnson(1954: 50) emphasizedan attractionbetween birds of the antwren flocks and thoseof the ant-followingaggregation. There seemsto be a social attraction,as the movingantwren alliance often joins or wheelsabout the more stationaryant-following alliance. However,I have noted little foraginguse of swarmsby antwrenseven in Amazonia, where there are more speciesof birds in antwren alliances than on Barro Colorado. Johnson stated that the antwrenor "social"aggregation "usually joins the feastingactivities" of the ant-followingbirds, but he refers mainly to Slaty Antshrikesand to Plain- brown Woodcreepersto support this contention. The latter are primarily ant-followers,and secondarilymembers of antwren alliances,rather than the reverseas he thought. The antshrikesoccasionally move down to the ants as the antwrensmove past, but usuallydrift off after a few minuteswith the ants. Some other medium to large species,such as Buff-throated Wood- creepers,also shift from one type of allianceto the other at times. Normally antwrensand most small membersof the antwrenalliance except migrants do not captureany prey flushedby the ants. If there are no ant-following birdsabout, birds of an antwrenalliance generally ignore an activeant swarm. Johnson(1954: 45) was the first ornithologistI know clearlyto distinguish betweenantwren and ant-followingalliances, but he did not presentdata to supporthis conclusionthat the wholeantwren alliance normally or actively forageswith ant-followingbirds as the alliancepasses them. 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 123

43 6O %

•.•.--_• 106 [] 70

4O v 5.5 / +38 " •+, / ,o,+ • •!3•;65 • /

2O +1961 •1962 YI963 = 964 ß 965 AI966

Jc•n Feb Mc•r Apr Mc•yJune July Aug Sept Oct Nov Dec

Figure 27. Percentageof Spotted Antbirds with the wandering antwren flocks when away from swarms of ants and nests.

ASSOCIATION WITH INTERSPECIFIC FLOCKS

Out of 798 Spotted Antbirds watched away from nests and away from swarms of ants from 1 November 1960 to 31 October 1961, some 299 or 37.5 percentwere with interspecificwandering flocks or joined interspecific flocks as I watched them. When I could not determine whether the bird waswith a flock, I did not countit in the 798. A SpottedAntbird was counted as beingwith a wanderingflock if one or more other speciesassociated with it for one minute or more. There is only a slightdecrease in the percentagesof SpottedAntbirds with flocks during the nestingseason, April to November (Figure 27). In some nestingmonths, such as August in 1964, Septemberin 1965, and June in 1966, the percentagesof SpottedAntbirds with flocks were over 40 percent. Perhapssome of the loweringof percentagesduring nestingmonths results from flushingSpotted Antbirds near undiscoverednests or young and from includingbirds leadingyoung about: birds limited in their movementsby nestsor youngare lesslikely to be with interspecificflocks, sincethese flocks move about rather rapidly at times. SpottedAntbirds readily join and follow wanderingflocks for varying 124 ORNITHOLOGICAL MONOGRAPHS NO. 10 periodsof time,but theyalways seem to be peripheralspecies. Their foraging methodstie them to watchingthe groundfor long periodsof time, so they followthe motionsof the moreactively foraging members of flocksvery im- precisely.Commonly a SpottedAntbird is left behindor driftsoff from the flock for brief periods. However,when it doeschange its foragingperch it often flies toward the flock rather than in other directions. SpottedAntbirds are especially likely to join flockswhen one watches them. Typicallythe antbirdmoves rather rapidly when one approaches,then settles downnear a densetreefall or near a flock of birds. Many SpottedAntbirds watched for five minutesor more ended up with bird flocks, even ones that first went to treefalls. Those that have seen the observermany times beforeare lessprone to moverapidly to treefallsor to flocks,however. Table 12 indicatesthat SpottedAntbirds with flocks tendedto preen or foragequietly somewhat more often than did birds awayfrom flocks,which tendedto mob,to panic,or to flee soI couldnot tell whatthey were doing. Of 749 observationperiods of SpottedAntbirds away from flocksfrom 1960 to 1966, 330 or 44.1 percentwere in the fleeing-mobbing-panickingcategories and 419 or 55.9 percentwere in the preening-foragingcategories. Of 608 recordsof SpottedAntbirds with flocks, 204 or 33.6 percent were in the former categories. Chi-squarefor these data is 15.52, highly significant (p <0.00008 accordingto Hartley and Pearson,1950). Some birds in the "alarmed"categories may have beennear an undiscoverednest or young; sincesuch birds tend to be awayfrom flocks,the totalsfor alarmedbirds away from flocks may be too high. However,in the mainly nonbreedingperiod from 1 Januaryto 30 April 1961, 34 of 144 SpottedAntbirds with flocks, or 23.6 percent,and 58 out of 132 awayfrom flocks,or 43.9 percent,were alarmed(Chi-square is 12.81, p < 0.00035). The first nine full monthsI worked on Barro Colorado,the percentageof alarmedbirds away from flocksexceeded the percentageof suchbirds with flocksby 12 to 47 percent.From 1 November1960 to 30 June 1961, 62 of 215 with flocks,or 28.8 percent,mobbed or panickedor disappeared;110 of 220 away from flocks, or 50 percent,were alarmed (Chi-squareis 25.09, p < 0.00001). From then until 1966, however,there have been 10 of 23 monthswhen the percentageof alarmedbirds was higherby 4 to 29 percent amongbirds with flocks(for all recordsfrom thento 1966 Chi-squareis 2.814, p < 0.1 and not significant). Probablythis happenedbecause, not realizing the significanceof alarm behavior,I tendedafter the first few monthsnot to record every chip or chirr. Also, my later work was censusing,with irregular coverageof tame birds someweeks and birds unaccustomedto the observerother weeks;and I persistentlyfollowed certain birds to see their bandswhether they were alarmed or not, and thusgot alarmedbirds habituated to me, rather than watchedany bird that was not alarmedas I did the first 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 125

TABLE 12 ACTIVITIES OF SPOTTED ANTBIRDS AWAY FROM SWARMS AND NESTS

First Year I 1961-1966 • Main Activity Number • Minutes• Number Minutes A. No young nor flock 1. Activity uncertain 38 131 42 235 2. Flees or chirrs 103 339 138 548 3. Wanders 76 440 92 497 4. Forages 181 2018 222 2173 5. Preens 5 32 10 49 6. Bathes 4 16 4 24 7. Feuds 84 734 81 735 B. No young; with flock 1. Activity uncertain 12 38 37 146 2. Flees or chirrs 76 259 75 342 3. Wanders 10 36 14 87 4. Forages 187 2087 192 2486 5. Preens 7 52 8 54 7. Feuds 23 113 31 316 C. Young; no flock 1. Activity uncertain 2 42 2. Flees or chirrs 27 601 42 820 3. Wanders 6 21 4. Forages 14 272 28 745 5. Preens 1 14 7. Feuds 12 79 D. Both young and flock 2. Flees or chirrs 24 285 4. Forages 3 60 33 564 • 1 November 1960 to 31 October 1961. • 1 November 1961 to 30 July 1966. ONum- ber of sessionswhen the activity occurred times number of birds performing it. • Two birds watched during one minute, "2 minutes" are recorded, etc. year. Tame birds,whether at or away from flocks,tend to forageand preen and ignorethe observer;to a certainextent birds in treefallsand near dense cover also are tame. Until I analyzedthe data and wrote the aboveaccount in early 1967, I had completelyoverlooked the tendencyof SpottedAntbirds to be more alarmed away from flocks. The above data thus came from a naive observer. In 1967 I watchedmore closely,and found that 43 of 71 birds away from flocks showedmore alarm behavior than foraging or preening (2 others showed equal time) while 10 of 75 birdswith flocksshowed more alarm behavior(1 more showedequal time); Chi-squarefor thesedata is 35.3 (p < 0.00001). 126 ORNITHOLOGICAL MONOGRAPHS NO. 10

In many casesflitting, flicking, darting to cover, and similar reactionswould not have beenrecorded unless I waswatching carefully. In somecases a bird chasedaway from a flock beganto chip or chirr and then stoppedalarm calls on reenteringthe flock. Less systematicobservations in 1968-1971 indicate that greateralarm outsideof flocks is normal behavior. Somebirds may join flocks becausemates can be locatedmore easily in flocks. Both sexesof SpottedAntbirds without matessing at or away from flocks, and court mainly at swarmsof ants where there is a good supplyof food for courtshipfeeding. They join flocks about as readily when mated as when mateless. When a bird comesoff the nest, its mate answersits songs as readily whetherit is with a flock or not, so it is not likely that the arriving bird normallyhomes on the flock. ForagingSpotted Antbirds with flocks generallywatch the groundrather than their companionsabove, so that their companionsdo not flush food for them. As is detailedbelow, they have differentforaging sites from all other speciesin the flocks; they do not go to food suppliesuncovered by other speciesor avoidplaces where other specieshave foragedany more than away from flocks. They foragemuch the sameway awayfrom flocksas in flocks. On a few occasions,Spotted Antbirds watchedScaly-throated Leafscrapers tossingdead leaves; once a leafscrapersupplanted such a watchingantbird. However,leafscrapers are rarely flock members.

OTHER MEMBERS OF FLOCKS Appendix1 listsfrequent members of wanderinginterspecific flocks of the forestinterior on Barro Colorado,with noteson their foragingbehavior and social structure and roles in flocks. There are several kinds of interspecific flocksnot consideredin this appendix,notably the bird flocks that follow army ants (see Johnson,1954: 41, and earlier parts of this paper) and the forest-edgetanager and honeycreeperflocks (Moynihan, 1962a: 2). Flocks of parrots and other birds in fruiting trees, overheadflocks of swifts, and wanderingflocks of toucansare alsooutside the boundsof this sectionof the presentpaper. The two typesof flocksconsidered are the groupsthat gather around White-flankedAntwrens in the forest undergrowthand the groups that gather around Gray-headedGreenlets in the upper levels of the forest. Using Moynihan'sterm "alliance"as shorthandfor a "wanderinginterspecific flock" or its separatedmembers, these may be calledthe "antwrenalliance" and the "greenletalliance," respectively. SpottedAntbirds attend only the antwrenalliances, since members of the greenletalliances forage high abovethem. However,the two alliancessometimes fuse as a large alliance,especially in placeswhere birds foraging at intermediate levels link the two. The antwren alliances on Barro Colorado are depauperatein species,especially in the open middle levels (15-20 m up), 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 127 and split apart from the greenletalliances more readily than do the species- rich and better-linked alliances in the Amazon basin and even those in other localities in the Canal Zone. Appendix 1 showsthe four main typesof birdsin thesealliances: dan- forming residents(probably mostly family groups); pair-formingresidents; solitaryresidents; and solitarymigrants. Two clan-forming(young stay with adults)residents, the White-flankedAntwrens and the Grey-headedGreenlets, are the "nuclear" species (in the sense of Winterbottom, 1943: 439, and not in the confusing[see below] sensesof later authors,including Winter- bottom, 1949: 259). The antwrenand the greenletare very similar in their foragingand socialbehavior, being ecologicalcounterparts for each other in the undergrowthand the forestcanopy. With Slaty Antshrikes,they are the commonest forest birds on Barro Colorado. Both look about rather activelyas they hop in the terminalsprays of greenfoliage, and both glean smallarthropods from the foliageor sally out to snap.them from the air nearby. Both are in territorial groupsof two to five birds most of the year, suggestingthat youngstay with their parentsup to a year as is likely for ant- tanagers(Willis, 1960a: 165) rather than beingdriven away as is the case for SpottedAntbirds. Membersof suchgroups keep in contactby frequent short chirpingnotes as they travel. A human approachingWhite-flanked Antwrenselicits loud "cheep doo" alarm calls and flight, but they soon becometame. The high-foraginggreenlets react with buzzingscolds when a human is nearby;normally they ignore the observeror arborealmammals after a few calls. Other forest-livingspecies of greenletsin the Amazon are mostlypaired residentsrather than clan-formingresidents, and follow rather than lead antwrenalliances; the alliancesseldom split into low and high- foraging groups in that region. The Dot-wingedAntwren, a regular member of forest flocks in British Honduras(Willis, 1960b), tendsto centeron densetangles of vinesat low to middlelevels in second-growthwoodland; it is becominguncommon on Barro Colorado as the forest matures, and its habitats are local, so that it tendsto form centersfor flocksin theseareas but to drop out of flocksas the other birdsmove into lesscluttered forest. In the vine tanglesand dense nearbyfoliage they flutter aboutactively in greenleaves for tiny insectsand are oftenseen in family-sizegroups that follow eachother with chirpingnotes and disputenoisily with othersuch groups. R. H. Wiley (MS) hasrecently confirmedstatistically that Dot-wingedAntwrens use densefoliage. The Dot-wingedAntwren has a chirr almostidentical to that of a SpottedAntbird when a humanfirst appears,but it very rapidly (comparedwith a Spotted Antbird) startsloud peupnotes or chirpsand resumesforaging. In the upper levels, the White-shoulderedTanagers are to some extent ecologicalcounterparts of Dot-wingedAntwrens. They forage actively on 128 ORNITHOLOGICAL MONOGRAPHS NO. 10 green foliage in vine-coveredtrees, they travel in small territorial groups, and they are becominguncommon on Barro Coloradoas the forest matures. They scoldloudly for a few momentswhen one approachesthem closelyat a forestoverlook but quicklyresume foraging and chirping. I am not sure whether they are more often leadersor followers of interspecificflocks, but on Barro Coloradotheir high mobility betweenpatches of habitat and their being uncommoninsure that they are usuallyattendants of the greenletal- liancesif they are with flocks at all. Sulphur-rumpedTanagers form constantlychirping family groupsor clans in the canopy. They forageboth in foliageand at fruiting trees,and move so rapidlythat they oftenleave greenlet flocks. At timesthey seemto be major speciesin greenletalliances, but at other timesthey join or form centersfor the forest-edge"tanager and honeycreeper"alliances. At timesthey link the two kindsof alliances,but at other timesthey wanderseparately. There are other such speciesin flocks elsewherein the Canal Zone. In humidforests of the Caribbeanslope of Panam•, but not on Barro Colorado, Tawny-crownedTanagers form noisy groupsin somealliances. SongWrens, which have disappearedon Barro Coloradoas the forest maturedbut form clannishgroups and sometimesjoin antwrenalliances in other woodedparts of the Canal Zone, are noisy when disturbedbut quickly habituateto the observeror elseflee. They differ from the other clan-formingbirds mentioned in poking under leaves on the forest floor and in similarly enclosedsites rather than foragingin open foliage. They commonlyassociate with Red- throatedAnt-Tanagers. I am not sure whetherSong Wrens follow or lead suchflocks, but they moveso slowlyit is surprisingthey often associatewith the motile White-flankedAntwrens and the even faster ant-tanagers. Red-throatedAnt-Tanagers, once commonon Barro Colorado but now becomingrare as the forest matures,and the Red-crownedAnt-Tanagers of drier woodlandsof the Pacific slope of the Canal Zone, sometimesform nuclear intraspecificgroups for interspecificflocks, but tend to move too rapidly to be permanentcenters. As a result, they tend to move from one flock to another,except at middaywhen their rate of travel is slow (Willis, 1960b). Many speciesof antwren and greenletalliances form pair bondsbut drive away their grown young and trespassers,so that one seesonly one or two birds of a speciesin each alliance. Such "pair-forming"birds tend to follow the nuclearor "clan-forming"species rather than lead them. On Barro Colo- rado,besides Spotted Antbirds, Checker-throated (Fulvous-bellied) Antwrens, Wedge-billedWoodcreepers, and Slaty Antshrikesare the commonestpair- forming members of alliances. Territorial pairs of Checker-throatedAntwrens chirp to each other; a human elicits loud peesk calls and flipping (pivoting from a positionfacing 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 129 right of the observerto a positionfacing left of him, or vice versa) back and forth, then flight; they becometame somewhatmore slowlythan do White- flankedAntwrens, but morerapidly than do SpottedAntbirds. Slud (1960: 100) suggestedthat Checker-throatedAntwrens forage in the same way as do White-flanked Antwrens, but I find that the former and their Amazonian relatives(the Brown-bellied,White-eyed, Rufous-tailed, and Stipple-throated antwrens) all forage in the sameway as North American Worm-eatingWar- blers;that is, all clamberlike furnariidsin green foliage and poke or rum- mage for arthropodsin scattered,dead rolled-up leaves that have not yet fallen from twigs in the forest undergrowth.They and their relativesfollow White-flanked or similar gleaning antwrens assiduously,and are regular membersof the forestflocks everywhere. The Wedge-billedWoodcreeper also occurssingly or in pairs, unlessde- pendentyoung are present,and follows antwrenalliances rather than leading them. It hitchesrather silently up tree trunks from near the ground to the forest midlevels,gleaning or prying small arthropodsfrom the bark and crevices. When antwrenscall loudly at a human or a potentialpredator, the woodcreeperadds a sneezingchi[! to the din. Like the antwrens,it fairly soon becomestame or ignoresthe observer. Slaty Antshrikesare so commonon Barro Coloradothat they are bound to be in flocksoccasionally; in addition,they oftenfollow flocksor are joined. Accordingto studiesby Yoshika Oniki (pers. comm.), pairs of Slaty Ant- shrikesoccupy small territories, drive away grownyoung or trespassers,forage by lookingabout carefully for largeto medium-sizedarthropods on twigsand in foliagefrom near the groundto 18 or 20 m up, and call faintly to each other as they move rather slowly about. Frequent dissectionof large prey and slownessof movementoften means that the antshrikesdrop behind a flock or follow its movementsimprecisely. In addition, they have such small territoriesthat a local antwrenalliance is often with a neighboringpair. Slaty Antshrikes are usually rather tame and phlegmatic in responseto humans, but can be incited to call ah, grrrrrrt, a caw followed by a growl, when ad- ditional factors (such as a nest or extremelynoisy antwrens) excite them. Dot-crownedAntvireos, relatively local and uncommonin areas of very openundergrowth on Barro Colorado,are intermediatebetween White-flanked Antwrensand Slaty Antshrikesin sizeand foragingbehavior but are like the latter in pair territoriality. They are usuallyfound in antwren alliances,but I am not familiar with their reactionsto humansor predatorsor their rolesas followers or leaders. The Plain Xenops,singly or in pairs,follows antwren flocks very persistently and quietly. It givesits wreep alarm call infrequently,mainly at hawks. It generallyignores humans or flees silently. It creepsor hangson rotten twigs in the middleand lower levelsof secondarywoodland, prying intently into the 130 ORNITHOLOGICAL MONOGRAPHS NO. I0 twigsor under the bark for tiny arthropods.Perhaps it is uncommonand local on my study area on Barro Colorado and in mature forestsgenerally becausethe twigs on which it feeds seldomlodge above the ground in the open and vertically-orientedlower levelsof tall forests. Like the Dot-winged Antwren, it is a regular memberof forest flocks in secondaryor woodland habitats, such as in British Honduras (Willis, 1960b) and in the Panamfi Canal Zone outside of Barro Colorado. Other intentlyrummaging furnariids are importantmembers of forestflocks, in most cases,either singlyor in pairs, in the highlandsof Panamfiand in Amazonianforests. One lowland species,the Buff-throatedFoliage-Gleaner, is a regular and persistentfollower in antwren alliancesin the Canal Zone but has disappearedfrom Barro Coloradoas the forest matured. It ignores the observerafter initial loud snare!calls, or flees silently. It rummagesin epiphytesor in piles of dead leavescaught in the undergrowthto midlevels of the forest;such piles of leavesand epiphytesdisappear from the lower midlevelsas horizontallimbs are shadedout and fall with forest growth. Two intently foraginglarge woodcreepersof the genusXiphorhynchus are regular membersof allianceson Barro Colorado, as are other membersof the genuselsewhere. Both occur singly or in territorial pairs, chaseoff tres- passersother than dependentyoung, and follow alliancesdesultorily and im- preciselyrather than leading them. Both hitch up tree trunks; they peer closelyand peck in crevicesor under crossinglianas, probe into epiphytes, and pry off loosepieces of bark with their long bills, but do not hammerwood or rummagedeeply into enclosedplaces. Black-stripedWoodcreepers tend to work the lower sidesof limbs in the canopyand midlevelsof tall forests; theyfollow greenlet alliances occasionally and antwrenalliances rarely. Buff- throated Woodcreeperstend to work the upper sidesof limbs in the middle and lower levels of woodlandsand secondgrowth. They follow antwren alliances strongly and greenlet alliances occasionally. The Black-striped Woodcreeperis commonon my studyarea on Barro Coloradowhile the Buff- throated Woodcreeperis uncommon;it is common and a regular flock- follower in woodlands elsewhere on Barro Colorado and in the Panamanian lowlandsgenerally. Both speciesgive loud alarmcalls and flee readilyon first sightinga humanor a potentialpredator; they are quiet and ignorea human later (i.e., they habituategradually). There are many other pair-formingmembers of flockson Barro Colorado, such as the Chestnut-backedAntbird of antwren alliances,the Long-billed Gnatwren of antwren alliancesnear densethickets or in secondgrowth, the TropicalGnatcatcher and Fulvous-ventedEuphonla of greenletalliances, and so on. All exceptthe Gnatcatchertend to forageintently or in denseplaces, all followrather than lead alliances,and all reactstrongly to potentialpredators and becometame gradually. 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 131

There are other flock membersthat are solitary,avoiding or chasingaway or ignoringmembers of their own species.These include one woodcreeper, severalflycatchers, and somemanakins. All forage openly,by flycatching or sallyinginto the air or to openfoliage. All flee readily,even to the extent of desertingflocks, even thoughmany seemrelatively tame in responseto a human at first. The Plain-brownWoodcreeper, a solitary and silent follower of flocks unlessa hawk or human causesit to give loud stiek calls, is uncommonin flocks mainly becauseit usually follows army ants. It is not tame, even after one follows a flock for some minutes, and it desertsreadily. It waits on tree and saplingtrunks in the upper undergrowthto the upper midlevels of the forest, then flies out to foliage or to trunks for the medium to large insectsit dissects.It followsflocks rather impreciselybut sometimesdoes so for long periods. Manakins,an importantfamily in secondgrowth and woodlandin tropical America, are poorly representedin forest flocks on Barro Colorado. All the speciestend to gatherat fruitingbushes and follow antwrenflocks rather irregularly--mainly when feeding on insects. Manakins are solitary to flockingbirds without pair or family organizationor evidentterritorial behavior, exceptat leks and when youngare followingfemales. They tend to ignorethe observer,then flee quietly, and they desertflocks readily. Ochre- bellied Flycatchers,which have becomeuncommon on Barro Colorado as the forest matured, behave as manakinsdo in most of these respects--as does the Plain-brown Woodcreeper,except that the woodcreepergoes to swarmsof ants rather than to fruiting trees. The Golden-collaredManakin, which salliesto the foliage of the low undergrowthin woodlandand second growth, has almostdisappeared on Barro Colorado as the forest grew tall and shadedout the denselower levels. The Red-cappedManakin, which sallies to foliage in the upper undergrowth,is now the only commonmanakin on Barro Colorado. The Red-cappedManakin fleessilently from the observerof flocks, althoughit becomestame at fruiting bushesor at leks. It probably takessome insects in competitionwith SpottedAntbirds when it forageslow, but generallythe two speciesforage at different levels. The Blue-crowned Manakin does well in Panamanianforests that do not have such a long dry season;it seemsto be unrecordedon Barro Colorado, but sometimesjoins alliances elsewhere in the Canal Zone. Tyrantflycatchers, members of thelargest family of tropicalAmerican birds, are also poorly representedin allianceson Barro Colorado. The solitary, silentBlack-tailed and Sulphur-rumpedflycatchers, which still follow antwren flocks and flit into the air in the upper understoryof woodlandselsewhere in Panamfi,have disappearedin the last 20 years as the forest matured. The minute Ruddy-tailedFlycatchers, solitary and silentbirds that follow flocks 132 ORNITHOLOGICAL MONOGRAPHS NO. 10 readily,are the only residentspecies still widelydistributed through the forest undergrowth.They generallyignore the observer,but flee silentlyif he stays nearby. They sallyshort distances to greenfoliage or into the air in the upper undergrowthof the forest,hunting very small insects.Bentbill Flycatchers, solitaryand silentunless displaying noisily at their dispersedleks, are small flycatchersof the low to intermediatelevels of the densevine tanglesfavored by Dot-wingedAntwrens. BentbillFlycatchers follow flocksimprecisely and are oftenseen away from them,because they stayin their patchesof habitat rather than with flocksthat move beyondthe patches.Olivaceous Flatbills, solitary and silent flycatchersof the undergrowthto midlevelsof irregular liana-crowdedforest in areasof treefalls,follow flocksreadily wheneverthe flock is near suchpatches of habitat but stay alone in their habitat at other times. The very small,generally solitary Golden-crowned Spadebills sally to foliagein the upperto lower undergrowthin areaswhere Dot-crowned Ant- vireoslive; they ignorehumans, are relativelysilent, and follow flocksrather rarely. Yellow-rnarginedFlycatchers sally to foliage in the midlevelsto lower canopyand follow greenlet alliances, as do Olivaceous,or Dusky-capped,Fly- catchers.Both speciesseem to travelin pairs,and to foragein rathervireolike ways,but moreobservations are needed.It may be that theseflycatchers are the counterpartsof the antvireosand antshrikesof the forestundergrowth. Largefrugivorous cotingas rarely join antwrenor greenletalliances, although in the Amazon and many other forests away from Barro Colorado small insectivorousbecards (Pachyramphus, Platypsaris) are pair-formingmembers of greenlet-antwrenalliances. Large woodpeckers,trogons (except insectiv- orousspecies of the forest undergrowth,such as Black-throatedTrogons on Barro Colorado), parrots,toucans (except small species),and other forest birdsseldom join alliancesfor verylong, althoughsome congregate at fruiting trees or go around together (toucans). Small insectivorousbirds are the main membersof forest alliances;even manakinsleave flocks when eating fruit. The replacementof oscines(songbirds) and membersof the superfamily Tyrannoidea(tyrant flycatchers,manakins) by the Furnarioidea(ovenbirds, woodcreepers,antbirds) in forest flocksis a generaltrend in the Neotropics as one goesfrom the forestcanopy or edgeto the forestundergrowth (Ap- pendix 1), from secondgrowth into deepforest, from savannaor dry areas into wet regions,from the highlandsto the lowlands,and from the periphery of the tropicsinto the heart of upper Amazonia. Antbirds tend to replace ovenbirdsand woodcreepers,also. On Barro Colorado, the growth of the forestin the 48 years sincethe islandwas set asidehas resultedin the loss of more songbirdsthan suboscines,in the virtual restrictionof residentsong- birdsto the canopyand forestedge, and in the disappearanceof many oscine 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 133 and tyrannidand dendrocolaptidand furnariidbirds from the antwrenalliances,as comparedwith suchalliances in lessmature forests a few hundred yards acrossthe PanamfiCanal. Antbirds are about the sameon the two sidesof the canalin numberof species,and are slightlymore abundant on the Barro Coloradoside in numberof individuals. They thus form a far higher proportionof the depauperateflocks on the Barro Coloradoside. Severalmigrant and winteringbirds, principallywarblers, readily join and follow antwren flocks on Barro Colorado and elsewhere in the Canal Zone. The three commonestspecies forage by flitting about in the foliagelike antwrens,but are solitaryand quietbirds. CanadaWarblers move about actively in thefoliage of the undergrowthfrom 1 to 15 m up. Chestnut-sidedWarblers are activein foliageof the middlelevels of the forest. Bay-breastedWarblers moveand peermore sluggishly at nearlyall levels,especially in the midlevels. Black-and-whiteWarblers creep over the trunks and large limbs of trees, pickingoff smallprey. AcadianFlycatchers snap up prey from the air and foliage of the lower levelsfrom 1 to 15 m up. All the migrantstend to be tame in responseto humans,to chip loudly or call excitedlywhen residentflock membersdo so, and to resumeforaging quickly after any interruption. None are very intent or extensiveforagers, althoughthe Acadian Flycatcherapproaches the latter condition and the Black-and~WhiteWarbler the former; they tend to be generalizedforagers at many levels. Thrushesand warblersthat forageintently (suchas the Wood Thrush and the waterthrushes,$eiurus spp.,) tend to be suchslow foragersthat they do not oftenfollow flocks;they tend to be extremelytimid and to haveprotective colors. Flycatchersthat forage extensively(such as wood pewees,Contopus spp.) tend to do so at local clearingsrather than move through the forest; they are seldomflock members. They ignore humansat first, but then disappear if one remains. One variably foraging winter resident,the Great Crested Flycatcher, sometimesjoins greenlet flocks in the canopy when foraging like a vireo; when it foragesextensively it is usually at clearing edgesand hencedoes not follow well. I am uncertainhow it interactswith the congenericDusky-capped Flycatcher.

COMPETITION AMONG FLOCKING BIRDS In contrastto the frequentsupplanting among flocks of birds over swarms of army ants, there is little overlap in foragingand little supplantingamong the birdsthat join wanderingflocks. The SpottedAntbird is especiallyisolated ecologicallyfrom other birds of the flocks, exceptfor the Chestnut-backed Antbird and the winteringKentucky Warbler, becauseonly thesethree species work the ground rather than the foliage or air as do most membersof the flocks. Occasionallya SpottedAntbird supplantsthe small White-flanked 134 ORNITHOLOGICAL MONOGRAPHS NO. 10 or Dot-winged or Checker-throatedantwrens when they forage near the ground,or the larger SlatyAntshrike supplants a SpottedAntbird when it movesup into the foliageor comesnear an antshrikedissecting prey on the ground,but most of theseother specieswork well above the groundeven whenthe SpottedAntbird is not present.Canada Warblers occasionally take prey like that takenby SpottedAntbirds, but the antbirdsand warblersgen- erallyforage in differentways and places. Chestnut-backedAntbirds, which are about the same size as Bicolored Antbirds,sometimes ignore the smallerSpotted Antbirds but at other times supplantthem with loud raspingcharngh! notes from distancesof as much as 20 m. I notedabove some other cases of supplantingson the infrequent occasionswhen Chestnut-backedAntbirds followed army ants, and an in- stancewhen one supplanteda SpottedAntbird at the nest. SpottedAntbirds normallyavoid Chestnut-backedAntbirds as they move aroundthe tangled treefallsor the densepatches of wild pineapplesthat the latter favors. In suchtangles the Chestnut-backedAntbirds forage mainly by hoppingnear or on theground and pecking or jumpingupward at moderatelylarge arthropods on overhanginglianas or leavesor trash;they shouldcompete with the open- foraging,downward-leaping Spotted Antbirds rather litfie. They oftencapture and dissectlarge prey, while SpottedAntbirds seldom do so. However, there is a considerableoverlap in the placesand ways of foragingand sizesand kinds of food taken. All recordedattacks have come since 1965, perhapsbecause Chestnut- backedAntbirds became very commonin densenew growtharound tangled treefalls after a windstorm on 1 October 1961. Since Chestnut-backed Ant- birdstend to stayin densetangles or movemainly between such tangles, they do not follow wanderingflocks persistently.Thus, they are only a minor and perhapstemporary exception to the rule that birds of the wandering flock do not attack each other. Of the winteringbirds that join flocks,only KentuckyWarblers forage much like SpottedAntbirds. Thesewarblers hop along the forest floor and over low lianasand debris,pecking here and there or hop-flutteringupward to peck their minute prey off low overhangingleaves (Willis, 1966a: 208). They thustend to forageupward for very smallprey, where there are sprouts, while the SpottedAntbirds forage downward in more open situationsfor larger prey, but there is some overlap. Away from swarmsI never saw antbirdssupplant warblers, becaue the lattergenerally keep out of the way of the similarlysized but heavierantbirds. However, Kentucky Warblers often hop after or aroundSpotted Antbirds, keeping at a distanceof 10 to 20 m. I suspectthat the Kentucky Warblers, which often forage at ant swarms, associatewith the SpottedAntbirds becausethe latter often lead them to ants. Oncea KentuckyWarbler cameto the taperecorder when I playedthe song 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 135 of a Bicolored Antbird, another swarm-followingspecies. The Kentucky Warbler does not follow antwren alliancesdosely or frequently, perhaps becauseit is somewhatslow-moving. One seldomsees supplantings of or by SpottedAntbirds away from swarms. In the years 1960-1966, I recordedonly 85 supplantingsand displacingsof SpottedAntbirds by othersof theirown species.In additionto the supplantings notedabove, I sawone BicoloredAntbird comeup to a male SpottedAntbird singingaway from a swarmand supplanthim three times,looking down from each successiveperch as if for ants. One other BicoloredAntbird supplanted a SpottedAntbird in another such incident. Three times on one occasiona SpottedAntbird supplanteda BentbillFlycatcher when it chirrednearby like a BicoloredAntbird (see Willis, 1967, Plate 1, for sonagramsof the very similarcalls of BentbillFlycatchers and BicoloredAntbirds.)

DISCUSSION The wanderinginterspecific flocks of tropicalforests resemble the chickadee- titmouse-warblerflocks of northern woodlandsin being composedof small insectivorousbirds that travel moderatelyrapidly throughthe forest. Northern flocksbreak up duringthe nestingseason to a greaterextent than do tropical flocks, which are evident most of the year. Probably one reason Spotted Antbirds and other membersof tropical flocks can stay with flocks during the nestingseason is that tropical birds have small broods and need not visit the nest as frequentlyas do northern birds. Conversely,the possible necessityof staying with flocks may add to the reasonsfor having small broods. The large numberof speciesavailable in tropical forestsprobably increasesthe possibilitiesof joining and following other species. Large territory sizesbecause of low numbersof individualsfor most species(James Karr, pers.comm.) may alsofacilitate following for longer distances. Many interspecificflocks are mainly assemblagesat concentratedsources of food or someother environmentalresource. Birds at swarmsof army ants and in fruiting trees,nesting assemblages of seabirdson islands,and roosting blackbirdsin marshesare examplesof this type of interspecificflock. How- ever, even these groups nearly always congregatemore denselythan seems necessaryfor efficientutilization of the environmentalresource. At the other extreme, judging by the distributional patterns of feeding, environmental resourcesseem to be scatteredfor the antwren alliances. Even though such birds sometimesforage irregularly,carefully working one area and moving rapidly through others, the alliancessometimes circle about and work the latter areascarefully only a few minuteslater. Although detailedanalyses are needed,the flocksseem to work wide areasrather evenlyover the course of a few hoursor days. Local areas sometimesvary greatly in food or other resourceseven in 136 ORNITHOLOGICAL MONOGRAPHS NO. 10 continuous forests. Some flocks in eastern Asia (Stanford, 1947: 508; McClure, 1967: 149) apparentlyfollow set routesand even appearat set timesof day. However,there seem to havebeen no long-termstudies by ob- serversactually tracking habituated or undisturbedflocks instead of simply encounteringa flock at the sameplace or time repeatedly.It is thusdifficult to know if temporal or spatial localization of food or another resourceis forcingthese birds to associate.If thereis spatiallocalization but no temporal localization, the birds should still scatter over the route of the flock unless flockinghas some advantage. Dot-wingedAntwrens and BentbillFlycatchers and other birdsthat prefer vine tanglesdo tend to restricttheir activitiesto suchareas on Barro Colorado. However, the general rule seemsto be that the more restricted a bird is to certain areas,the less likely it is to be a consistentmember of antwren or greenletalliances. The Dot-wingedAntwren is a much more consistentflock member in British Honduras (Willis, 1960b) where its habitat is fairly widespread,than on Barro Colorado or in the Amazon where its habitat is restricted. The most consistentmembers of flocks are those birds, such as White-flankedAntwrens, that usethe habitatmoderately evenly. The specieswith scatteredhabitats are restrictedin manycases because they cannotfollow movingflocks. Dot-wingedAntwrens follow antwrenalliances persistentlywhen these alliances are in vine-crowdedareas but drop out as the alliancesmove into more open forest. SpottedAntbirds also follow antwren alliancesfairly persistentlywhen coveris nearby,but tend to drop out when the antwrensmove into very openundergrowth. Following and leadershipis clear in most of theseflocks, and is another line of evidencethat indicatesthe birds staytogether rather than beingforced togetherbecause of localizedresources. So far only Moynihan (1962a: 18) has studiedfollowing and leadingquantitatively, and he has worked with forest-edgeflocks rather than with flocks of the forest interior. However, preliminaryindications are that SpottedAntbirds and perhapsother members of the antwrenalliances are not forced togetherby localizedenvironmental resources,and that they do actively associatewith each other. The Spotted Antbird seemsto use a bird flock as it usesa treefall, as a safeplace to stop when one chases it. Active associationin an area of scattered resources,rather than even or randomscattering, would seemdisadvantageous to birds because(1) prey of one bird mightbe eatenor frightenedby other birds of the flock (Goss- Custard,1970: 18-19) (2) predatorsof one bird mightbe attractedby the noise,etc., of the flock (3) one speciesmight physicallyor competitively interferewith the activitiesof another,since (4) slowingdown or speedingup so flockscan staytogether is alwaysa disadvantage(Nichols, 1931: 181), and, since(5) stayingwith flockscan interferewith feedingyoung and with 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 137 otherreproductive activities, (6) reactivityto other speciesmay use DNA or brain cellsand energythat couldbe usedfor feedingor reproduction.If any of theseoccurs, or evenif they do not, thereshould also be someadvan- tagesin associatingwith other species. Thosewho have reported on flockshave suggested many possible advantages. The followinglist is probablyincomplete, and the categoriesoverlap, but it will do for a start. The earliestauthors or major authorsI have found for a suggestionare listed;"perhaps" means that the authorstates the sug- gestionindirectly or only hintsat it. A. Food-givingadvantages (1) Animalsflush food for eachother (perhapsBelt, 1874: 123; Neave, 1910: 80). (2) Animalsmay pilfer from othersor get food from their leavings (Rand, 1954: 31). (3) Animals lead each other to good food sources (Nichols, 1912: 45). (4) Animalscan avoid sitesjust usedby others(Miller, 1922: 125), or nichesusually used by others (Morse, 1967: 101). B. Predation-avoidingadvantages (5) Intently foraginganimals forage best if aeriallyalert foragers give the warning(perhaps Moynihan, 1962a: 120; Willis, this report). (6) The more animals,the more likely they are to seepredators; scareone, scareall (Bates, 1863: 347). (7) More animals allow individuals to hide behind each other, using each others' bodies as cover (Williams, 1964; W. D. Hamilton, fide Goss-Custard,1970: 35). (8) More animalsconfuse a predator(Miller, 1922:123; Lorenz, 1963: 142). (9) More animalscan mob, attack, collide with, or intimidatepred- ators; aggressiveanimals attract weaker ones (Swynnerton, 1915: 348); loud-voiced speciesprotect faint-voiced ones (R. H. Wiley, MS). (10) Clumpingincreases the irregularityof the environmentfor predators,making them move further for each try at prey, given that (5)-(9) above occur; reducedpredator success lowers predator numbers (Trivers, 1971: 44). C. Mate-gainingadvantages (11 ) Animalsget mates more easily by associating. (12) Birdslocate or staywith matesor youngmore easily (Willis, this report). 138 ORNITHOLOGICAL MONOGRAPHS NO. 10

D. Stimulationadvantages (13) Speciesfacing occasionaltimes of loweredstimulation, such as drab winter forests,may need "socialfacilitation" to bring them to normal or necessarylevels of food-searching(L. Kil- ham, pers.comm.) or predatoravoidance (Moynihan, 1962a: 120). E. Learning advantages (14) Young, migrant, vagrant,or otherwiseinexperienced birds can learn local predatorsor sites where predatorsoccur or seldom occur. (15) Inexperiencedbirds can learn where there are locally superabundant foods. (16) Inexperiencedbirds can learn how others forage and avoid competition. (17) Inexperiencedbirds can learn where to go for various resources,such as a lake or winteringground or roostsite. F. Navigation advantages (18) Averaging variable headingsof individualsgives correct re- suitantto home (Hamilton, 1967: 58). G. Population-controladvantages (19) Animals can estimateinterspecific competition and restrictre- productiverates (Wynne-Edwards,1962: 418).

What few data there are suggestthat the SpottedAntbird and most other speciesof Neotropicalforest-interior flocks gain few food advantagesby associating.I rarely saw one speciesflush food for anotheror locatea good food sourcethat anotherspecies then used. It was neverevident that Spotted Antbirds or other speciesavoided foraging in the samesites more than would be the caseby chancealone. Miller (1922) suggestedthe latter from watching bushtits(Psaltriparus minimus), but Richard B. Root (pers. comm.) has seenbushtits foraging one after the other in the samefoliage. Statistically valid studiesare neededfor bushtitsand other flockingspecies. Several lines of evidencesuggest that food advantagesare usually not primary in the Neotropicalalliances. First, flycatchingbirds are not common in such alliances,whereas they shouldbe the main membersif birds flush prey for each other. Tyrant flycatchers,although the largestfamily of Neo- tropicalbirds, are mostlyuncommon in flocks. The yellow-rumpedflycatchers of the genusMyiobius, the Dusky-cappedFlycatcher, and the flycatching CanadaWarbler are exceptions,and may well exploit food flushedby other specieseven if most speciesof the alliancesdo not. Swynnerton(1915) mentionssavanna flocks that definitely gathered around drongosflushing food; I have notedbrief associationof SpottedAntbirds with Scaly-throated 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 139

Leafscrapers.These casesgrade into flocking associationssuch as those around army ants, and may play varying roles in many flocks. A secondline of evidencethat food advantagesare not primary in the antwren alliances is that most such birds take different foods in different ways (R. H. Wiley, MS, has recentlyconfirmed this quantitativelyfor three speciesof antwrensin flockson Barro ColoradoIsland), and fight with each other or go for the sameprey itemsrelatively infrequently. These behavior patternscontrast strongly with thoseamong groups where food is the primary reason for social aggregation,as it is for birds over swarms of army ants. Separationof nichesis the rule amongbirds of antwren alliances,overlap of nichesthe rule amongbirds with ant-followingflocks. The relationsof Spotted Antbirdswith competitorsat swarmsof antsand in antwrenalliances illustrate the differencebetween the typesof flocksvery well, sincethe samespecies is involved. More data are neededto judge if some speciesoverlap, since Morse (1967, 1970) suggeststhat someflocking birds separate their foraging nichesbetter than do lone birds, presumablyby avoidingniches of other speciesor becauseof aggression. Third, northernbirds seldom join interspecificflocks when they are feeding young,at the time their food needsshould be highest. Fourth, SpottedAnt- birds and manakinsand other birds that gatherat concentratedfood sources commonlyleave alliancesto do so. Fifth, birds that find abundantfoods in localizedhabitats (includingmost of the ant-followingbirds that do not forage away from ants) generallystay in suchhabitats or move directly from one to the other rather than wanderingwith alliances. There is alsolittle directdata for or againstthe theorythat predationis lesson birds that join flocks. Rudebeck(1950: 87) found that European Sparrow-hawks(Accipiter nisus) catchmore prey outsideflocks and succeed on more attemptsoutside flocks, but Morse's (1970: 163) analysisindicates the differenceRudebeck found in Sparrow-hawksuccess is not significant.Of course,the figuresshould be in termsof successfulattempts per 100 birds in and outsideflocks, for it may be that a predatoronly makesattempts on flocks whenit is ascertain of gettingprey as whenit attacksan individualbird. That is, if the Sparrow-hawkavoids attempts on birds in flocks there may be an advantagefor prey to flock eventhough the few attackson unpreparedflocks (such as notedby Tinbergen,1946: 96-97, for the cross-hedgerowstyle of hunting)are relativelysuccessful. Even the easilyobserved raptors of open countryoften turn out to have unexpectedforaging methods (see Willis, 1963, for one example);and data are almostlacking for raptorsin tropical forests.However, radio transmittersmay soonmake it possibleto get more data, so that studentsof raptorsmay be able to get quantitativeobservations 140 ORNITHOLOGICAL MONOGRAPHS NO. 10 on kills per 100 birds on attemptsinside or outsideof flocks, as well as information on how hawks react to flocks. Unfortunatelyfor quantitativestudies, bird-eating forest-falconsof the genusMicrastur, hawks of the genusAccipiter, and owls of the genusGlau- cidium are typicallybirds of secondgrowth and forest edges,not of the ex- tensivetracts of forest that are best for interspecificflocks. Bird-eating hawksseem to do best at interfacesbetween habitats, or in irregular habitats; the Sparrow-hawksthat pounce on birds by crossinga dense hedge (Tin- bergen, 1946: 96-97) are an example. There are many speciesof hawks in tropicalforests, and mostterrify smallbirds or catcha few, but the majority of Neotropicalraptors specializeon snakesand other animals that do not form interspecificor intraspecificflocks. One possiblespecialist on birdsof forestflocks is the Tiny Hawk, Accipiter superciliosus,a bird so small (20-28 cm long) that it apparentlycan approach interspecificflocks without being detected--at least, two of the four timesI have seen it it darted through flocks unsuccessfullyso fast that the birds scarcelyhad time to react. (Perhapstoucans and large insectivorouscotingas and trogonscannot join antwren alliancesbecause the small birds flee from them?) Even the Tiny Hawk worksthe canopyand forestedge, not the forest interior. Barred Forest-Falconsalso work ant-followingflocks and antwren alliancesto someextent, but do so by scatteringthe flock and then waiting for regroupingbirds or (at ant swarms)feeding on insects.They stayin secondgrowth woodland. The negativecorrelation between bird-eating hawks and antwren alliances couldbe consideredone point in favor of the theorythat birds join flocksto escapepredation, but the correlationcould be causedby concurrentlyvarying factors: lesslight in the forest interior, for instance. A secondpoint is that somebirds are less nervousin flocks than outside of flocks. Murton (1967) reportsthis for Wood Pigeons(Columba palumbus) in England. Lack (1968: 135) suggests,apparently from his own casualob- servations,that birds in flocks can feed more efficiently becausethey waste lesstime lookingaround nervously. The data for SpottedAntbirds point this way. Moynihan (1962a: 120) reportsthat Plain-coloredTanagers are lessshy towardsman away from flocks than in flocks. His commentthat "tame and unsuspicious"birds may be warnedby other more alert or suspiciousspecies almoststates the hypothesisthat intently foragingbirds join alert speciesthat call the alarm (seefollowing paragraphs) but fails to suggestan evolutionary advantagefor being unwary away from flocks. It seemsthat the level of shynessshould be directlyproportional to the danger,so that a specieswhich is not shy away from flocks and is shy within them is either in more danger 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 141 when it joins flocks or is wastingforaging time looking for predatorsand beingalarmed when the dangeris to otherspecies and not to it. Under either condition,I wouldexpect the speciesto cometo avoidflocks. In some cases,it may be adaptivefor birds to associatewith others to increasetheir own vigilance. If a bird normallylives a predator-freelife and only occasionallygoes into predator-richzones, it may then be best for it to join othersso it can follow the old dictum"When in Rome, do as the Romans do." Presumablyits level of vigilancewill be adaptivefor the old zone and not for the new or infrequentzones in such cases. The bird may also learn local predatorsin suchcases. Moynihan (pers.comm.) has found that non- flocking birds isolatedfrom their normal habitatsoften join flocks. The tendencyof migratorybirds to join flocks is probably due to learning advantagesof this and other types. However,I doubt that the residentPlain- coloredTanagers were shyerin flocks than outsideflocks for thesereasons. Moynihan'stanagers probably had little to fear from him. Treetop birds like thesegenerally lack special"chirring" or raspingcalls for groundpredators, and small birds usually have little to fear from large, slow-moving animalslike humans. Moreover, the tanagersaround the Barro Colorado clearingwhere Moynihan worked have long lives (Crebbs, 1964) and have becomeespecially habituated to the numerousscientists. Their greater ner- vousnessin flocksis probablya maladaptiveresponse to alarm notesor movements of other speciesthat have less experiencewith man, not an adaptive response.In other words, if Moynihan had been a significantpredator he couldhave pouncedon tame tanagersaway from flocksand eliminatedthem from the evolutionarypicture, but it is doubtful that doing so would make tanagerstame and unsuspiciousaway from flocks. PossiblyMoynihan meant that Plain-coloredTanagers were "alarmed" rather than "shy," as the former usually implies displayingand calling at dangerrather than fleeingfrom it. Alarm behavioris highestat intermediate levelsof dangerrather than directlyproportional to danger;shyness or flight replacesalarm behaviorat high levels,and other activitiessuggesting lack of concernat low levels. Birds inexperiencedwith man in antwren alliances certainlyact as if man werea significantpredator when he first appears,and call the alarmloudly. After the initial callsand displays,however, birds are generallyless shy or alarmedthan if they are alone. Solitarybirds in flocks, like lone onesaway from flocks,tend to ignorea humanor call, then vanish. Nuclear speciesof flocks, onesthat forage moderatelyintensively and form clansof theirown species, tend to call loudlyand then forage without showing alarm behavior. Pair-formingmembers of flocks tend to call loudly and becometame rather slowly, but do not desertflocks the way solitarymembers often do (Appendix 1). 142 ORNITHOLOGICAL MONOGRAPHS NO. 10

A third point that tendsto supportthe hypothesisthat birds join flocks to escapepredation is that SpottedAntbirds and many other speciesthat follow alliancesare birds that forage intently (i.e., examinenearby and en- closingsurfaces carefully), while birds that form intraspecificor nuclear groupsfor alliancestend to be onesthat forageby looking more extensively (i.e., checkingfoliage and air at moderatedistances). Birds that look very extensively(i.e., checkingfor prey at considerabledistances and lookingabout slowly) suchas flycatchers,tend to be solitary and either ignore flocks or desertthem readily (Appendix 1). Migrant warblers,which forage moderately extensivelyand join flocks, perhapsdo so rather than form family groups becausethey scatterduring migration. Also, as noted above,they can learn local predatorsor otherlocal conditionsfrom local birds. Perhapsrummagers and birds working in enclosedsites or looking at the groundor tree trunks would have to waste much foragingtime watchingfor predatorsif they did not join speciesthat can be alert for aerial predatorsat the same time as they forage.* There may thusbe a definitevalue in associatingwith other speciesrather than with other membersof the same species,especially where there are many specializedspecies that forage intently but need to move rapidly. Six Checker-throatedAntwrens, all stickingtheir headsin rolled-upleaves, would probablybe less safe than two Checker-throatedAntwrens associating with four White-flankedAntwrens flitting in the open foliage. The hypothesis suggestsone possiblereason wh3' the White-flankedAntwren keepsits grown youngwith it whilethe Checker-throatedAntwren doesnot, eventhough both are in the samegenus. It also explainswhy the SpottedAntbird and other intent foragersgenerally drive grown youngaway while many medium-intent foragerskeep their youngwith them: competingyoung are valuableif they can help look about for predatorsbut not so valuableif they cannot. Fly- catchingbirds would gain little by keepingtheir young or even mates,both becausethey needto work large spacesand would competewith eachother and becauseeach individual can keep its own lookoutrather well. Leafscrapersand waterthrushesare intentforagers but are exceptionsto the rule that suchforagers tend to follow flocks. Leafscrapersprobably forage too slowlyto be able to follow flocks. Hence they are solitaryand have de- velopedvery cryptic plumagesand inconspicuousbehavior patterns as an alternateway to avoidpredation. Both they and waterthrushesflee if flushed, giving loud notes. Waterthrushesare sometimeslimited to stream banks,

* William Dilger (letter) suggeststhat specieslike Hermit Thrushes, which forage in enclosingvegetation, tend to look about more rapidly than specieslike Eastern Bluebirds, which watch for prey and predators in the open. Perhaps the rapid looking about of many species of enclosed vegetation puts them in greater danger from predators than does slow and inconspicuousscanning of many birds that look very extensively. 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 143 but other reasonsfor their not followingflocks are not evident. They are protectivelycolored. Some birds, such as Red-throatedAnt-tanagers, move too fast (Willis, 1960b) to be good alliancemembers; such birds tend to developflocks of their own species.Such birds can scarcelybe intentforagers unless predators are few. Probablythe developmentof a foragingniche is limited by the developmentof behaviorpatterns and morphologysuch that the nichewill not causetoo high a death rate for the possiblereproductive rate. Following an interspecificflock is oneway of exploitinga nichethat requires considerable activity without incurring excessivepredation. Following is possibleonly within a certain range of speedsand environmentswhere the nuclear or followedor "host" specieshave the right behaviorpatterns. If the various speciesadjust their speedsto each other, as has been noted for ducksand shorebirds(Nichols, 1931: 181) in flying groups,they may have to forage lessefficiently or feedtheir youngless well but can survivebetter. Swynnerton(1915: 354) found that somebirds associatewith noisy, ag- gressivedrongos. Similarly, R. H. Wiley (MS) suggeststhat on Barro Colo- rado an intent forager, the Checker-throatedAntwren, has a call that is more effectivein mobbingthan thoseof the two medium-intentspecies of antwrens that forage with it. The conspicuouscall, possiblebecause the Checker- throatedAntwren is protectivelycolored (a necessityfor an intent forager on hangingdead leaves) and perhapsimportant in its own protection,may help and thus attractother speciesthat do not have suchcalls. In this case, intent speciesgain by havingless intent specieskeep the lookout while less intent speciesgain by havinga bird aboutthat is able to mob predatorsmore effectively;this wouldbe a symbioticassociation. The SpottedAntbird and many other intent foragershave similarlyloud calls, which may warn other speciesof a flock or help them mob predators as well as being of value to the individualbird itself. However, the Spotted Antbird and Checker-throatedAntwren and other specieswith such sharp calls seemto join flocks rather than lead them. Since the main advantagein flockingshould be to the followers,the sharpnessof the mobbingcall is unlikely to be as important as the advantageto an intent forager of having nearby birds that can look for predatorsall the time while foraging. More data on who follows whom are needed to evaluate the relative contribution of the loud calls and of the protectionof intent foragersto flock organization. Conspicuousmobbing calls also characterize certain important members of Amazonianflocks, especiallythe Cinereousand Dusky-throatedantshrikes. Birdsmay join thembecause they are both goodat keepingthe lookout (being "flycatchers") and at mobbing. The mate-findingfunction of joining interspecificflocks is uncertain. SpottedAntbirds generallysing for mates,and do so whetherthey are with 144 ORNITHOLOGICAL MONOGRAPHS NO. 10 other speciesor not. It may be that singingis somewhatinhibited when the bird is not with a flock, but comparativedata are lacking; this would be another antipredationadvantage. Much of courtship feeding occurs when birds are at swarmsof ants rather than in flocks, but the fact that less time is used in fear behaviorbecause of joining a flock could make more time availablefor foragingor courting. Possiblysome birds find the mate more easilywhen leaving the nestif the mateis with a flock, but SpottedAntbirds seemto singto each other rather than home on antwrenalliances. Sinceonly one bird out of five would be with an alliancerather than with ants or away from both, going to allianceswould be somewhatinefficient. However, I found it usefulto checkall antwrenalliances when doing a censussearch for particular individual Spotted Antbirds, so Spotted Antbirds may use the same techniqueat times. The possibleadvantages of learningby joining antwrenor other alliances have not been demonstrated.If predatorsare not specialists,learning which onesto avoid wouldcertainly be possiblein alliancesat lesscost to a species than in intraspecificflocks or clans. It is possiblethat Spotted Antbirds learnthat certainparts of the forestare safeby stayingwith antwrenalliances; but the safe places for the low-dwellingSpotted Antbird are unlikely to correspondto safe sitesfor the higher-dwellingmajority of the flock. A bird may alsolearn what kinds of placesare exploitedby other speciesand avoid wastingtime investigatingsuch places; but for the SpottedAntbird there are more competitorsthat do not join antwren alliances than there are ones that do. SpottedAntbirds live in a rather homogeneoushabitat, henceprobably do not need to join a flock to navigateto a food site or safe "home" or stay within it; in fact, joining an interspecificflock is more likely to lead a bird away from home. The possibilitythat SpottedAntbirds may assessinter- specificcompetition and then adjust reproductiverates by joining antwren alliancesis also remote, becausethey competelittle with other membersof alliances. The peripheraland irregularattendance of SpottedAntbirds with antwren alliancesprobably results partly from their slowforaging and partly from their hesitation at moving into the more open sites that alliance members go through. Moreover, when food is available near treefalls or other sources of cover the SpottedAntbird can use theseas substitutesfor the protective flocks. SpottedAntbirds are probablymainly followersbecause they forage intently and inconspicuously.They gain little by forming clans, and thus are quietmuch of the time they are in flocks(notes to the mate or youngare infrequent);other speciescould not follow them easilyeven if there were a reason to follow. Few largebirds occur in antwrenalliances; the largestmoderately frequent 1972 WILLIS: BEHAVIOR OF SPOTrED ANTBIRDS 145 memberis the Black-throatedTrogon (20-25 cm long). Perhapslarge size frightensother birds. Somelarge birds move too slowly (woodpeckers)or too rapidly (toucans), and many othersare fruit eatersand hencelimited by fruiting trees. Fruit-eatingspecies sometimes form other types of alliances, someof whichwander in somewhatthe sameway as do the antwrenalliances (seeMoynihan, 1962a: 2); at othertimes the alliancesat fruitingtrees may be competitiveor internallyantisocial aggregations like thoseover army antsbut still have somepredator-avoiding functions. The antwrenalliances and the ant-swarmflocks, both importantto Spotted Antbirds,are at oppositeextremes of the spectrumof tropicalflocks. On BarroColorado the "blueand green tanager and honeycreeper" flocks (Moyni- han, 1962a: 2) are almost intermediatein characters--thereis a moderate amountof interspecificcompetition and fighting and a moderateamount of socialwandering and stoppingto feed. Moynihanclassflies the birds in flockshe studiedas "active" versus "passive" species,the formerfollowing and the latter beingfollowed. However,by this terminologythe relativelyinactive Spotted Antbird is an "active"species and the constantlymoving White-flankedAntwren a "passive"species. Since intent foragerswill generallybe "active"followers and flitting speciesoften will be "passive"leaders, this is generallytrue. I suggestthat the words "following" and "followed" or "attending" versus "attended" are self- explanatoryand less confusing. The terms "nuclear"and "circumference"species have had a confusing historysince Winterbottom (1943: 439) proposedthem for what I call the clan-formingand solitaryto pair-formingbirds. Althoughhis termssuggest centralversus peripheral location in flocks, he was trying to indicatethat intraspecificgroups form nuclei for attendant birds. Davis (1946: 169) dividedthe speciesof flocksin Brazil into "regular"or usuallyflocking and "accidental"or rarely flocking. Winterbottom(1949: 259) then buried the idea of intraspecificflocks as nucleiby combininghis and Davis'scate- goriesinto a confusingsystem: nucleus,other regular, regular accidental, and accidentalspecies. Moynihan (1962a: 67), reviewingthe confusion,decided to call "nucleus" speciesthose that contributeto flock formationwhether they are followers or leaders,while "attendant" species were those that "do muchless to stimulate the formation and/or maintain the cohesionof mixed flocks." Like Winter- bottom,he suggestedthat intraspecificflocks evolutionarily are often suitable centersof interspecificflocks; but he suggestedthat manyother species later becomeintegral parts of flocks. Moynihan also used the terms "regular" and "occasional"(he considered "accidental"misleading) for the percentagesjoining or associatingwith flocks: he specificallystates that a rare speciesthat regularlyfollows flocks may 146 ORNITHOLOGICAL MONOGRAPHS NO. I0 statisticallybe lessfrequent in themthan a commonspecies that is occasional in flocks. McClure (1967: 146) attributed other meaningsto Moynihan's terms, those of relative abundancein flocks: "speciesusually presentwere regular,less often seenones occasional." Confusionabounds, and somemay feel it bestto drop termsthat arbitrarily split intergradingphenomena until we have more data on the phenomena themselves.Perhaps we should either drop the confusingterms "nucleus" and "attendant"and "regular" and "occasional"or replace them with de- scriptive, self-explanatoryadjectives. "Nucleus" speciesand "attendant" speciesare deeplyimbedded in the literature,to be sure. But, wherethese mean "elan-forming"and "followers,"respectively, in Winterbottom's(1943: 439) terminology,they refer to dozensof charactersin Moynihan's terminology. There they may refer to noisy versusquiet species,to drab versusbright species,to aggressiveversus retiring species,etc. It seemsmore direct to use the appropriateadjectives or to decideon one meaningfor "nuclear." For McClure's (1967: 146) definitionsof "regular" versus"occasional," we can speak of "frequent" and "infrequent" species,if we must break a series into two parts; for Moynihan'sdefinitions of "regular" versus"occasional," similarly,we can speakof "persistent"and "nonpersistent"species. It is possiblethat elan-formingbirds are lessimportant as followed species in someflocks in some regions,especially in forestsin the Amazon Basin. Most flock membersthere seemto be pair-formingbirds rather than dan- forming birds, even the antwrens(Gray Antwren, Long-wingedAntwren) that divide up the foragingniche of White-flankedAntwrens or occur with them over much of the region. Evolution shouldbe toward greater speciali- zationand towardgreater intensity of foragingin species-richfaunas. "Jacks- of-all-trades"ordinarily become less common in suchfaunas, as parts of their nichesare successivelyexpropriated by competitionfrom evolvingspecialists. The greaterintensity of foragingand greaterspecialization in rich avifaunas shouldresult in pair-formingbirds replacingdan-forming ones. Not only is there too little food in a givenarea to allow youngto be toleratednear their parents,but the youngcannot help parentsmuch in predatordetection if young are foragingvery intently. In rich faunaslike thoseof the Amazon, speciesthat form pairs may group directly and follow each other alternately,or follow noisyand conspicuous species, rather than follow clans as do birdsin Panamfi. The family or elan as a unit of organizationseems to be most frequent amongbirds of secondgrowth or woodland,especially in areasof moderate speciesdiversity; pair-forming species take over in tall forests,with higher diversity,as well as in open areas with low diversity. Red-crownedAnt- tanagers,mostly pair-forming birds of forests,and Red-throatedAnt-tanagers, mostly elan-formingbirds of secondgrowth, illustrate the shift within a singlegenus (Willis, 1960a: 165). Selander(1964: 206) notesthat elan 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 147 formation in cactuswrens is characteristicof mesic environments,as is clan formationin the genusThryothorus. In wrens generally,forest and desert speciesare paired or solitary while speciesof intermediatehabitats form clans. Factorsfavoring clans in successionallyintermediate types of habitats areperhaps high reproductive success (Snow and Snow, 1963: 40); moderately low speciesdiversity, with the consequentnecessity of flocking with one's own speciesif there is any value to flocking;irregular or uncertainfood suppliesthat may be superabundant,if they occur at all, and hencecan be exploitedby all of a family; etc. The hypothesisthat there is a decreasein clan-formingspecies as one enters either forest or "wasteland"from such intermediatehabitats as canopy, forest edge, secondgrowth, or savanna shouldbe checkedstatistically, of course. Suchtrends are known for monkeys, ungulates,cats, and other animals;I do not know if they occur in humans, unlessthe city can be considereda forestfor humans. Moderateirregularity of the environmentwith respectto predatorsand food supplies,rather than absolutelevels of thesefactors, is probably the basic factor leading to the prevalenceof clan formation over the formation of interspecificor occu- pationalgroupings. This is becauseclans are lesscommon in desertsand other "wasteland"areas where conditions are extremelyand uniformlyharsh as well as in situationswhere physical conditionsare extremely and uniformly favorable. 148 ORNITHOLOGICAL MONOGRAPHS NO. 10

SUMMARY The small, sexuallydimorphic Spotted Antbird (Hylophylax naevioides, Formicariidae)joins other birds at swarmsof army ants and in interspecific flocks in lowland forests between Honduras and Ecuador. A detailed ethologicalanalysis, mainly accomplishedby observationof color-banded birdson Barro ColoradoIsland in the PanamfiCanal Zone, suggestshow its behavioris related to its habitat, to competitionfor food, and to avoiding predators. Preeningand other maintenanceactivities take little time, but avoiding suchsmall environmentalhazards as sunlight,rain, and externalparasites accountfor somebehavior patterns, perhaps even mutual grooming in mated pairs. Freezingas a reactionto potentialpredators and other large environmental hazardsis rare; there is no specializedcall ("keening") with the postureas thereis in somerelated antbirds (Willis, 1967:13). PerhapsSpotted Antbirds forageso activelythat freezingwould seldombe usefulin avoidingpredators. Far more commonis hyperactivepanicking and chipping,which may serve four functions: warningor teachingrelatives, warning or startlingpredators, confusingpredators, and scaring dominant competitors. Spotted Antbirds mob and chirr at passinghumans or other mammals,as do related antbirds; thenoise and display perhaps causes predatory mammals to moveaway because any prey would be alerted. Spotted Antbirds are relatively incurious and slow to becometame, perhapsbecause they are slow fliers and forage in moderatelyopen undergrowth and would not be safeif they were tame or came to squeakingnoises. Perhapssubmissive displays, "whimpering" calls and "cringing," are rare becausebirds usually flee rather than stay to face a dominant bird. Two aggressivedisplays are more common: "high-challenging"is a brief and introductoryupright posture associated with a "bugling"call; long "snarling" callsgo with "low-challenging,"a similarbut horizontalposture that displays the paleback-patch and breastconspicuously. Agonistic displays are strongest at the centersof territoriesand weakest at the boundaries;the theorythat these displaysarise from conflictsof attackand escape "drives" seems less universally applicablethan the theorythat aggressivedisplay arises from interferencewith attackand submissivedisplay from interferencewith flight in situationswhere the opponentis about the samesize. (When the "opponent"is small, maintenancebehavior is used;when the "opponent"is large, reactionsto danger are used.) Courtshipbehavior involvesfive main activities: matelessmales sing loudly;mateless females wander to them;chirping and "flirting" occurwhen a bird of the oppositesex appears;courtship feeding by the male cements 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 149 the pair bond and precedesall copulafionsand nestings;mutual grooming occursamong mated birds and betweenparents and young. Agonisticbehavior is not prominent in courtship;perhaps it is useful mainly for speciesof ephemeraland otherwiseirregular habitats, where there are shortpair bonds and many sibling species. Young males setfie on territoriesin their first year. Pairing occurswhen the male is on territory. A femaleusually first nestsin the breedingseason after her hatching. Females,perhaps because they are subordinateto trespassingmales, usuallywander nomadicallyto new mates if the old ones disappear. Widowed males, which are still wholly dominant on their own territories, generally stay on them. The nestcup is usuallypendent from slendertwigs on a smallsapling, 0.3 to 1.4 m abovethe ground. Building,by both male and female,takes several mornings. Material is gatherednear the nest. Two eggsare laid, two days apart. Incubation,performed by both sexes,required 15 daysin two cases and 16 daysin one;the two youngat one nesthatched a few hoursapart, but the two youngat the last nestnearly a day apart. Male and femalefeed young, which stay in the nestalmost 12 days. The completenesting cycle takes 35 to 40 days. Young can barelyhop whenthey leavethe nest. Parentslead the youngto low perchesin densevegetation and perform striking distractiondisplays if one tries to capturethe young. The male of a pair feedsone fledgling,the female feeds the other. As the young acquire adult plumage, complete by aboutsix weeksafter leavingthe nest,the parentsstop feeding them and finallydrive them off. Parentsrenest soon after or beforethe previousbrood becomesindependent. Predators rob over 90 percent of nests,but pairs renestup to 10 timesduring each rainy season,April to November. The help by the male at all stagesof nestingis probablyrelated to habitat, becauseanimals of regular habitatsare generallysexually egalitarian while animals of irregular habitats tend to have division of labor and sexual dimorphismand diethism. Independentyoung show little alarm behaviorand little agonisticbehavior when they wander nomadically,perhaps becauseany noise would attract dominantadults. The changefrom the "age rule" of dominanceto the "territorial rule," after the youngbird getsa territory,marks the point wheresimple aggressionstops and aggressionbraked by limits begins;territoriality is not simpleaggression but a limit on aggression. A SpottedAntbird follows swarmsof army ants about half the time. It movesto the peripheryof swarmsor to a positionhigh abovethe ants when largerantbirds attack it. When excludedfrom the large and regularnomadic raids of Eciton burchelli,it finds the irregularswarms of Labiduspraedator, irregularstatary raids of Eciton burchelli,or foragesaway from ants al- 150 ORNITHOLOGICAL MONOGRAPHS NO. 10 together. Althougha subordinatespecies, it has many foragingalternatives. At one pair per 4.7 hectares(12 acres), this generalizedspecies is much commonerthan the larger and dominantcompetitors that are more restricted to feeding with army ants. A SpottedAntbird foragesmainly by waitingnear the groundand darting down to snapup small prey, of one beak lengthor less. It prefersperches of small diameter but often uses vertical perchesunless it is preening. At swarmsof ants it movesmore activelyand attemptsto captureprey about four timesas rapidly as it doesaway from swarms.If competitorsare present, it indulgesin warblerlikeactivity, wanderingand snappingup prey. Away from ants it sometimesjoins the wanderinginterspecific flocks of antwrens and other birds ("antwren alliances") of the forest interior. It is somewhatmore timid with respectto humanswhen away from the alliances, and flees either to suchflocks or to densecover when chased. Perhapssuch intently foraging speciesas the Spotted Antbird join interspecificflocks becausethey would otherwisehave to use more foragingtime showingalarm or lookingfor predators. This hypothesisand observationssuggest that birds that tend to forage activelyin sucha way that they cannot be alert are followers,unless they hide in densecover or are very cryptic;that speciesthat forageby lookingabout moderately alertly form intraspecificgroups and are leaders;and that birds (like flycatchers)that look aboutvery alertlyto forage tend to be solitaryand to desertflocks readily. For birdson Barro Colorado, thereis little evidencefor and severalpoints against the alternativehypotheses that birds of a wanderingalliance flush food for eachother, lead eachother to food, excludeeach other from their own niches,locate mates, or practice populationcontrol by evaluatinginterspecific competition (Wynne-Edwards, 1962: 418). Speciesthat join suchflocks are onesthat rarely disputeor take the same foods or niches,for instance. Many of the behavioraland morphologicalcharacteristics of SpottedAnt- birdscan be correlatedwith the low degreeof irregularityof their environment in regard to food and predation. A moderatelyhigh degreeof aggressionat army ant swarmsis probably an adaptation to the variable niche of a subordinatespecies; the definitebut poorly enforcedterritorial systemis an aggression-limitingadaptation to a more predictableenvironment away from swarms. Aggressionis perhapsmost useful when food suppliesor covercan be defendedand are worth defending,which is most likely to be true in irregular environments.Sexual dimorphismand female flirting or submission are probably also adaptationsto moderatelyirregular lives when Spotted Antbirds are excluded by dominant large birds at ant swarms; otherwise SpottedAntbirds are sexuallyegalitarian, as is usually true of animals of regularhabitats. Much evidencefor birdsindicates that moderatelyirregular environmentsare an influencein the other direction: that is, toward strong 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 151 courtshipdisplays, aggressiveness, strong male dominance, short pair bonds, polygamyor promiscuity,one sex only caringfor offspring,and alert or crypticsolitary foraging. Intraspecific dan formation also occurs in moderately irregularenvironments, however. Similar trends have been noted in human societiesof economicallyirregular environments--the slum societies studied by OscarLewis (1968) andthe Bedouinsstudied by JohnGlubb (1963), for instance. GAZETTEER OF LOCALITIES Localitieswhere I have observedSpotted Antbirds are listed here. Coordinatesrefer to approximatesites of observations,which may be severalkilometers from the town or geographicalfeature used as a name. Coordinatesare to the nearestminute of north latitude and west longitude,respectively. Coordinates are followedby elevations, in meters,estimated from availablemaps or from altimeterreadings, and by the average yearlyrainfall, in millimeters,estimated from the vegetationand from averagesat nearby stations. PANAM•- Cerro Campana.--8ø 40', 80ø 04'; 900 m; 2,700 mm. Mountaintop,partly forested and partly cleared,overlooking Punta Chamewest of the Canal Zone. Antbirdsseen 29 August 1961 and 25 to 28 June 1968. Cerro Azul.--9 ø 15', 79ø 20'; 700 m; 2,500 mm. Ridge, partly forested,northeast of TocumenAirport and east of the Canal Zone; observations23 June 1964, 31 May 1966. PANAMA CANAL ZONE Agua Salud.--9ø 12', 79ø 48'; 25-100 m; 3,000 mm. Creek into Gatun Lake north of Frijoles, in wet forest about 40 to 60 years old. Barro Colorado.--9ø 10', 79ø 51'; 25-165 m; 2,730 mm. Forestedisland in Gatun Lake (see text). BohioPeninsula.--9 ø 12', 79ø 51'; 25-100 m; 3,000 mm. Peninsulainto Gatun Lake north of Barro Colorado, with secondaryforest 40-60 years old. BuenavistaPoint.--9 ø 11', 79ø 50'; 25-50 m; 2,800 mm. Peninsulainto Gatun Lake northeastof Barro Colorado; secondaryforest 20 to 40 years old. EscobalRoad.--9 ø 14•, 79ø 58'; 110 m; 3,500 mm. Rolling hills, wet forest40-60 yearsold, on sideroad by Rio Medio north of Gatun Lake. MaddenReserve.--9 ø 06•, 79ø 37'; 50-200 m; 2,500 mm. Secondaryforest 10 to 60 years old on rolling hills along continental divide.

COLOMBIA Apartad6.--7ø 56', 76ø 40'; 100 m; 3,000 mm. Patchesof wet lowland forest on flats northeast of town; visited 8 to 9 March 1965. Chigorod6.--7ø 45', 76ø 40'; 130 m; 3,500 mm. Patchesof wet foreston flatsnortheast of town and river; visited 10 March 1965. Puerto Belgica.--7ø 43', 75ø 17'. 130 m; 2,500 mm. Isolatedpatches of tall forest in pasturesnorthwest of town on CaucaRiver; visited9-10 June 1962. Remedios.--7ø 02', 74ø 41'; 770 m; 2,500 mm. Extensiveforests north of town, on rolling hills; SpottedAntbird seen5 May 1962. Rio Verde.--A branch of the Rio Sinfi, in partly cutover hill forests east of the Serranla de Abibe. Rainfall from 2,500 mm. at mouth of Verde (7 ø 505 76 ø 17') at 150 m elevation to over 4,000 mm. on west slope of Filo de Abibe (7 ø 45 •, 76 ø 31') at 152 ORNITHOLOGICAL MONOGRAPHS NO. I0

635 m. Except on the east slope of the Filo on 26 March, SpottedAntbirds seen every day on hike up and back, 20 to 28 March 1965. San Pedro.--8 ø 27', 76 ø 18'; 150 m; 2,500 mm. Patchesof dry forestson rolling hills west of Rio San Juan and north for five kilometers, 12 and 15 March 1965. Tanand6.--5 ø 37', 76 ø 39'; 60 m; 9,000 mm. Very wet, nearly flooded low forestsjust westof Rio Atrato a few miles upstreamfrom Quibd6; 23 February 1962. Tucurgt.--7 ø 56', 76 ø I0'; 150 m; 2,800 mm. Wet forests on rolling hills east of Rio Sinfi, 16-17 June 1962; forestscut down by 1965. Yuto.--5 ø 30', 76 ø 32'; I00 m; 9,000 mm. Very wet hill forest by road south to Rio San Juan, a few kilometers from the Rio Atrato; 24 February 1962.

LITERATURE CITED

AKERMAN,B. 1966a. Behavioural effects of electrical stimulation in the forebrain of the pigeon. I. reproductive behaviour. Behaviour, 26: 323-338. AKERMAN, B. 1966b. Behavioural effects of electrical stimulation in the forebrain of the pigeon. II. protectivebehaviour. Behaviour,26: 339-350. ALLEE, W. C. 1926. Measurement of environmental factors in the rain-forest of Panama. Ecology, 7: 273-302. ARMSTRONG,E. A. 1949. Diversionary display. Pt. II. The nature and origin of distraction display. Ibis, 91: 179-188. BASTOCK,M. 1967. Courtship: an ethologicalstudy. Aldine, Chicago, 220 pp. BATES,H.W. 1863. The naturaliston the river Amazons. Murray, London, 394 pp. BELT, T.W. 1874. The naturalistin Nicaragua. Longmans,London, 403 pp. BERITOFF (BERITASHVILI), J. S. 1965. Neural mechanisms of higher vertebrate behavior. Little, Brown, and Co., Boston,384 pp. BLURTONJONES, N. G. 1968. Observations and experiments on causation of threat displaysof the Great Tit (Parus major). Animal Behav. Monographs, I: 73-158. BOWMAN,R.I. 1961. Morphological differentiation and adaptation in the Gal•tpagos finches. Univ. California Publs. Zool., 58: 1-326. BROWN,J. L., AND R. W. HUNSPERGER.1963. Neuroethology and the motivation of agonisticbehavior. Animal Behav., 11: 439-448. C'I-IAmN,J.P. 1932. The birds of the Belgian Congo, Part I. Bull. Amer. Mus. Nat. Hist., 65: 1-729. CREBBS,T.C. 1964. Longevity in the Plain-coloredTanager, Tangara inornata. Bird- Banding, 35: 264. CROOK,J.H. 1964. The evolution of social organisation and visual communication in the weaver birds (Ploceinae). Behav. Suppl., 10. CULLEN,J.M. 1963. Allo-, auto-, and heteropreening. Ibis, 105: 121. DAvis, D.E. 1946. A seasonal analysis of mixed flocks of birds in Brazil. Ecology, 27: 168-181. EISENMANN,E. 1952. Annotated list of birds of Barro Colorado Island, PanamgtCanal Zone. Smiths. Misc. Colls., 117(5): 1-62. GLUBB, J. 1963. The empire of the Arabs. Prentice-Hall, Inc., Englewood Cliffs, N.J., 384 pp. GOss-CusTARD,J. D. 1970. Feeding dispersion in some overwintering wading birds. Pp. 3-35 in Social Behaviour in Birds and Mammals (J. H. Crook, editor). Academic Press, London, 492 pp. HAMILTON,E. J., III. 1967. Social aspectsof bird orientation mechanisms.Pp. 57-71 in Animal Orientation and Navigation (R. M. Storm, editor). Corvallis, Oregon State University Press, 134 pp. 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 153

HARTLEY,H. O., nNOE. S. PEARSON.1950. Tables of the xMntegral and of the cumu- lative Poisson distribution. Biometrika, 37: 313-325. HEmaER,H. 1950. Wild animals in captivity. Butterworth and Co., London, 207 pp. HINGE, R.A. 1961. Behavior. Pp. 373-411 in Biology and Comparative Physiology of Birds, II (A. J. Marshall, ed.). Academic Press,New York, 468 pp. HINGE, R.A. 1969. Animal behaviour, Second Edition. McGraw-Hill, New York, 876 pp. HOLST, E. VON, ANo U. VON SAINT PAUL. 1963. On the functional organization of drives. Animal Behav., 11: 1-20. HUTCHINSON,G.E. 1951. Copepodologyfor the ornithologist.Ecology, 32: 571-577. IERSEL,J. J. A. VAN, nN• A. C. A. BOL. 1958. Preening of two tern species. A study on displacementactivities. Behaviour, 13: 1-88. JEHL,J. R., JR. 1970. Sexual selectionfor size differencesin two speciesof sandpipers. Evolution, 24: 311-319. JOHNSON,R. A. 1954. The behavior of birds attending army ant raids on Barro Colorado Island, Panamfi Canal Zone. Proc. Linn. Soc. New York, 63-65: 41-70. KAUFMANN,J.H. 1962. Ecology and social behavior of the coati, Nasua narica, on Barro Colorado Island, Panama. Univ. California Publs. Zool., 60: 95-222. KELso, L., ANDM. M. NICE. 1963. A Russian contribution to anting and feather mites. Wilson Bull., 75: 23-26. KRUIJT,J.P. 1964. Ontogenyof socialbehaviour in Burmesered junglefowl (Gallus gallus spadiceusBonaterre). Behaviour Suppl., 12. LACK,D. 1943. The life of the Robin. Witherby, London. LACK,D. 1947-48. The significanceof clutch size. Ibis, 89: 302-52, 90: 25-45. LAClC,D. 1968. Ecological adaptations for breeding in birds. Methuen, London, 409 pp. LEHRMAN,D.S. 1958. Induction of broodinessby participationin courtshipand nest- building in the ring dove (Streptopeliarisoria). J. Comp. Physiol. Psychol.,51: 32-36. LEWIS, O. 1968. A study of slum culture, backgroundsfor La Vida. Random House, New York, 240 pp. LORENZ,K. 1963. On aggression.Harcourt, Brace and World, New York, 306 pp. MACARTHUr,R. H., •a,•oE. R. PIANKA. 1966. On optimal use of a patchy environment. Amer. Nat., 100: 603-609. McCL}•¾, R. A., AN• R. Y. MOORE. 1965. Subcortical mechanismsof behavior. Basic Books, Inc., New York. McCLum•, H.E. 1967. The compositionof mixed speciesflocks in lowland and sub- montane forests of Malaya. Wilson Bull., 79: 131-154. MARGALEF,R. 1968. Perspectivesin ecological theory. Univ. Chicago Press,Chicago, 111 pp. M•LER, P. 1955. Characteristicsof some animal calls. Nature, 176: 6-8. MA•LER, P. 1956. Behaviour of the chaffinch (Fringilla coelebs). Behaviour Suppl., 5: 1-184. MARLER,P., AND W. J. HAMILTON, III. 1966. Mechanisms of animal behavior. John Wiley and Sons, New York, 771 pp. MA•IELO, H. 1961. Nesting successcalculated from exposure. Wilson Bull., 73: 255-261. MEYER DE SCHnUENSEE,R. 1966. The speciesof birds of South America and their distribution.Livingston, Narberth, Pennsylvania,578 pp. MILLER, A.H. 1964. JosephGrinnell. Syst. Zool., 13: 235-242. MILLER,R.C. 1922. The significanceof the gregarioushabit. Ecology,3: 122-126. 154 ORNITHOLOGICAL MONOGRAPHS NO. 10

MOREA•J,R.E. 1960. Conspectusand classificationof the Ploceine weaver-birds. Ibis, 102: 298-321, 443-471. Momus, D. J. 1957. "Typical intensity" and its relationship to the problem of ritualization. Behaviour, 11: 1-12. MORSE,D. H. 1967. Foraging relationships of Brown-headed Nuthatches and Pine Warblers. Ecology, 48: 94-103. MORSE, D. H. 1970. Ecological aspects of some mixed-speciesforaging flocks of birds. Ecol. Monogr., 40: 119-168. MOYNIHAN,M. 1955. Some aspectsof reproductive behaviour in the black-headed gull (Larus r. ridibundusL.) and related species.Behaviour Suppl., 4. MOYNIHAN,M. 1962a. The organization and probable evolution of some mixed speciesflocks of neotropical birds. Smith. Misc. Colls., 143(7): 1-140. MOYNIItAN,M. 1962b. Display patterns of tropical American "nine-primaried"songbirds. I. Chlorospingus.Auk, 79: 310-344. MURTON, R. K. 1967. The significance of endocrine stress in population control. Ibis, 109: 622-623. NEAVE, S.A. 1910. On the birds of northern Rhodesia and the Katanga district of Congoland--Part I. Ibis, 9(4): 78-155. NICE, M.M. 1943. Studiesin the life history of the song sparrow, II. Trans. Linn. Soc. New York, 6: 1-328. NICHOLS,J.T. 1912. Notes on recognition marks in certain speciesof birds. Auk, 29: 44-48. NICHOLS,J.T. 1931. Notes on the flockingof shorebirds.Auk, 48: 181-185. ONIrd, Y. 1970. Nesting behavior of Reddish Hermits (Phaethornis ruber) and occurrenceof wasp cellsin nests. Auk, 87: 720-728. OaIaNS,G.H. 1969a. The studyof life. Allyn and Bacon,Inc., Boston. OmeNS,G. H. 1969b. On the evolution of mating systemsin birds and mammals. Amer. Nat., 103: 589-603. OmeNS,G. H., aN•) G. W. C}mISTMAN.1968. A comparativestudy of the behavior of Red-winged,Tricolored, and Yellow-headedBlackbirds. Univ. California Publs. Zool., 84: 1-85. PAINE,R.T. 1966. Food web complexityand speciesdiversity. Amer. Nat., 100: 65-75. PEARSON,D.L. 1971. Verticalstratification of birdsin a tropicaldry forest. Condor, 73: 46-55. POTTEa,E.F. 1970. Antingin wild birds,its frequencyand probablepurpose. Auk, 87: 692-713. Ram), A.L. 1954. Socialfeeding behavior of birds. Fieldiana: Zoology,36: 1-71. RETTENMEYER,C.W. 1963. Behavioralstudies of army ants. Univ. Kans.Sci. Bull., 44: 281-465. RICKLEFS,R.E. 1968. Weightrecession in nestlingbirds. Auk, 85: 30-35. RICKLEFS,R.E. 1969a. An analysisof nestingmortality in birds. Smiths. Contrib. to Zool., 9: 1-48. RICKLEFS,R. E. 1969b. The nesting cycle of songbirdsin tropical and temperate regions. Living Bird, 8: 165-175. RI?LE¾,S. D. 1959. Competitionbetween sunbird and honeyeaterspecies in the Moluccan Islands. Amer. Nat., 93: 127-132. ROWELL,C. H.F. 1961. Displacementgrooming in the chaffinch. Anim. Beh., 9: 38-63. 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 155

RUDEBECK,G. 1950. The choiceof prey and modesof huntingof predatorybirds with specialreference to their selectiveeffect. Oikos, 2: 65-88. SCHNEmLA,T. C. 1957. A comparisonof speciesand generain the ant subfamily Dorylinaewith respectto functionalpattern. InsectesSociaux, 4: 259-298. ScoTt, T. G. (Editor). 1963. Grouse managementsymposium. J. Wildl. Mgrnt., 27: 525-887. SELnNDER,R. K. 1964. Speciationin wrens of the genus Campylorhynchus.Univ. California Publs. Zool., 74: 1-305. SELAND•R,R.K. 1966. Sexualdimorphism and differentialniche utilization in birds. Condor, 68: 113-151. SHORT,L. L., JR. 1961. Interspeciesflocking of birds of montane forest in Oaxaca, M6xico. Wilson Bull., 73: 341-347. SIBL•¾,C.G. 1957. The evolutionaryand taxonomicsignificance of sexualdimorphism and hybridizationin birds. Condor, 59: 166-191. SKtrrcH,A.F. 1946. Life historiesof two Panamanianantbirds. Condor, 48: 16-28. SKUTCH,A.F. 1949. Do tropical birds rear as many young as they can nourish? Ibis, 91: 430-455. SKUTCH,A.F. 1952. On the hourof layingand hatchingof birds'eggs. Ibis, 94: 49-61. SictrrcH,A.F. 1962. The constancyof incubation.Wilson Bull., 74: 115-152. SictrrcH,A. F. 1969. Life historiesof Central American birds, III. Pacific Coast Avifauna, 35:580 pp. SLUD,P. 1960. The birds of Finca "La Selva,"Costa Rica: a tropicalwet forest locality. Bull. Amer. Mus. Nat. Hist., 121: 49-148. SLim,P. 1964. The birdsof Costa Rica. Bull. Amer. Mus. Nat. Hist., 128: 1-430. SMITH,N.G. 1969. Provokedrelease of mobbing--a huntingtechnique of Micrastur falcons. Ibis, 111: 241-242. SMXTI-I,W.J. 1966. Communicationsand relationships in the genusTyrannus. Nuttall Orn. Club Publ., 6: 1-250. SNow,D. W., n•D B. K. S•ow. 1963. Breedingand the annual cycle in threeTrinidad thrushes. Wilson Bull., 75: 27-41. STn•FORD,J.K. 1947. Bird partiesin forestsin Burma. Ibis, 89: 507-509. STOKES,A.W. 1962. The comparativeethology of great,blue, marsh,and coal tits at a winter feeding station. Behaviour, 19: 208-218. SUTTON,G. M. 1968. Reviewof "The Shorebirdsof North America,"G. D. Stout (ed.). Wilson Bull., 80: 500-503. SWX•N•RTO•,C. F.M. 1915. Mixed bird parties. Ibis, 3: 346-354. TI•B•ROE•,L. 1946. Die Sperwerals Roofvijandvan Zangvogels(with Englishsum- mary). Ardea, 34: 1-213. Ti•a•RO•, N. 1940. Die 13bersprungbewegungen.Zeits. f. Tierpsych.,4: 1-40. TI•a•RO•, N. 1959. Comparativestudies of the behaviorof gulls (Laridae): a progressreport. Behaviour, 15: 1-70. TINaERO•N,N. 1963. The shellmenace. Nat. Hist., 72(7): 28-35. TROY,RS,R.L. 1971. The evolutionof reciprocalaltruism. Q. Rev. Biol.,46: 35-57. V•R•R, J., n•D M. F. WILLSO•. 1966. The influenceof habitatson matingsystems of North Americanpasserine birds. Ecology,47: 143-147. WHITAKER,L.M. 1957. A r6sum6of anting,with particularreference to a captive orchard oriole. Wilson Bull., 69: 195-262. WILLIAMS,E.C. 1941. An ecologicalstudy of the floor faunaof the Panamarain- forest. Bull. Chicago Acad. Sci., 6: 62-124. WILLIAMS,G.C. 1964. Measurementof consociationamong fishes and commentson the evolutionof schooling.Michigan State Univ. Mus. Pub., Biol. Ser., 2: 349-384. 156 ORNITHOLOGICAL MONOGRAPHS NO. 10

WILLIS, E. 1960a. A study of the foraging behavior of two speciesof ant-tanagers. Auk, 77: 150-170. WmLIS, E. 1960b. Red-crowned ant-tanagers, tawny-crowned greenlets, and forest flocks. Wilson Bull., 72: 105-106. WILLIS, E. 1961. A study of nesting ant-tanagers in British Honduras. Condor, 63: 479-503. WILLIS E.O. 1963. Is the zone-tailed hawk a mimic of the turkey vulture? Condor, 65: 313-317. WILLIS E.O. 1966a. The role of migrant birds at swarms of army ants. Living Bird, 5: 187-231. WILLIS E.O. 1966b. Interspecific competition and the foraging behavior of plain- brown woodcreepers. Ecology, 47: 667-672. WILLIS E.O. 1967. The behavior of bicolored antbirds. Univ. California PubIs. Zool., 79: 1-132. WILLIS E. O. 1968. Studies of the behavior of Lunulated and Salvin's antbirds. Condor, 70: 128-148. WmLIS E.O. 1969a. On the behavior of five speciesof Rhegmatorhina,ant-following antbirds of the Amazon basin. Wilson Bull., 81: 363-395. WILLIS, E.O. 1969b. Review of A. F. Skutch's "Life Histories of Central American Highland Birds." Auk, 86: 363-365. WINTERBOI•FOM,J.M. 1943. On woodland bird parties in northern Rhodesia. Ibis, 85: 437-442. WINTERBOTTOM,J.M. 1949. Mixed bird parties in the tropics,with specialreference to northern Rhodesia. Auk, 66: 258-263. WOoD-GusH, D. G. M. 1956. The agonistic and courtship behavior of the brown leghorn cock. Br. J. Animal Behav., 4: 133-142. WOODreNO,W.P. 1958. Geology of Barro Colorado Island, Canal Zone. Smiths. Misc. Colls., 135(3): 1-39 WYNNE-EDWARDS,V. C. 1962. Animal dispersion in relation to social behaviour. Oliver and Boyd, Edinburgh, 693 pp.

APPENDIX 1 CHARACTERISTICSOF MEMBERS OF FORESTFLOCKS ON BARROCOLORADO ISLAND

Characteristic Species a b c d e f g h i j k 1 m n o

Antwren, White-flanked 1 1 1 1 1 1 1 1 c c c a a 1 1 Greenlet, Gray-headed 1 1 1 1 1 1 1 1 c c c g g 5 gl Tanager, Sulphur-rumped 1 1 1 1 2 u u c g g 5 g8 Tanager,White-shouldered 1 1 1 2 u u u g g 5 g6 Antwren,Dot-winged 1 1 1 1 1 1 1+ 2 u u u a a 1 9 Gnatcatcher,Tropical 1+ 1 2 2 2 1 u u r g g 5 g4 Euphonia,Fulvous-vented 1+ 1 2 2 2 2 u c c g g 5 g7 Antvireo, Dot-crowned 1+ 1 2 2 2 2 2 1 u u c- a+ a 1 11 Antshrike, Slaty 2- 1+ 2 2 2 1+ 2- 2 c c c a* a 1 3 Gnatwren,Long-billed 2- 1 2 2 2 2- 2 1 r r r a a 5 19 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 157

A??V.NDIX 1 (continued)

Characteristic

Species a b c d e f g h i j k 1 m n o

Antwren, Checker-throated 2 2 2 2 2 2 2 1 c c c a a 1 2 Antbird, Spotted 2 2 2 2 2 2 2 2 u u c a- a I 8 Antbird, Chestnut-backed 2 2-2 2 2 2 2 2 r u u a- a 1 18 Flycatcher, Dusky-capped 2+ 1- 2 2 2 2 u c c g g 4 g3 Flycatcher, Yellow-margined 2* 1 2- 2 2 2 c c c g g 4 g2 Xenops, Plain 2 2 2 2 2 2 2 1 u u r a a 2 10 Woodcreeper, Wedge-billed 2 2 2 2 2 2 2 1 c c c a a 2 4 Woodcreeper,Buff-throated 2 2 2 2 2 2 2 1 c u r a a 2 7 Woodcreeper, Black-striped 2 2 2 2 2 2 2 2 u c c g g 2 g5 Puffbird, White-whiskered 3- 2+ 2 3 3 3- 2 2 u u c+ a a - 12 Flycatcher, Ruddy-tailed 3- 1- 3 3 3 3-2 1 c c c a+ a 4 5 Flycatcher, Bentbill 3- 3+ 3 3 3 3- 2 2 u u u a a 4 13 Flycatcher, Ochre-bellied 3- 3+ 3 3 3 3- 2 2 u u r a a 4 16 Spadebill,Golden-crowned 3- 3+ 3 3 3 3- 2 2 u u c a a 4 17 Flatbill, Olivaceous 3 3 3 3 3-3-2 I u u u a a 4 14 Manakin, Red-capped 3 3 3 3 3 3 2 2 c c c a a 3 6 Manakin, Golden-collared 3 3 3 3 3-3-2 2 r r r a a 3 20 Woodcreeper,Plain-brown 3 3 3 3 3 3 2 2 u r c a+ a 2 15 Redstart, American I- 1- 3 3 3 3 2 1 u u c+ g g 5 m9 Warbler, Canada I- 1- 3 3 3 3 2 1 c c c a a 5 ml Warbler, Chestnut-sided 1- 1 3 3 3 3 2 1 c c c a a 5 m3 Warbler, Bay-breasted 1+ I 3 3 3 3 2 I c c c a+ a 5 m2 Vireo, Red-eyed 2- 1 3 3 3 - 2 1 u u r g g 5 m8 Warbler, Kentucky 2- 1+ 3 3 2 3 2 2 u u c a a 5 m7 Warbler, Black-and-white 2 2 3 3 3 3 2 1 u u c g- a 5 m6 Flycatcher, Great Crested 2+ 1- 3 3 3 - 2 2 u u c g g 4 m5 Flycatcher, Acadian 2+ I- 3 3 3 3 2 2 u c c a a 4 m4 a--Foraging method: 1, moderatelyintent; 2, very intent; 3, lookswidely. b--Food source: 1, greenfoliage; 2, dead material; 3, both and air. c•Family structure: I, clansformed; 2, pairs; 3, solitary. d--Chirping notes: 1, frequent; 2, occasional;3, rare. e--Initial alarm: I, noisy;2, sometimesnoisy; 3, quiet. f---Subsequenttameness: 1, rapidly tame; 2, slowly tame; 3, tendsto flee. g--Leader?: 1, leadsor is followed;2, follows. h--Persistence: I, high per cent with flocks; 2, moderateor low. i--Frequency in flocks: c, common; u, uncommon;r, rare. j--General abundance on Barro Colorado: c, common; u, uncommon; r, rare. k--Forest type used: c, forest; u, vines and tangles;r, secondgrowth. 1--Foraging height: a, low; g, high. m--Flock type joined: a, antwren; g, greenlet. n Taxon: I, antbirds; 2, furnariids; 3, manaklns; 4, flycatchers; 5, songbirds. o--Rank in Barro Colorado alliances: 1-20, antwren alliances;gl-g8, greenlet alliances; ml to m9, migrants. + or- means tending higher or lower; in many casesI is higher than 3 and 3 is lower than 1, in a closed triangular series. APPENDIX 2 x

•tccipiter spp.--see Hawks Antbird, Bicolored ( Gyrnnopithysbicolor)--4ff. Dot-backed ( Hylophylax punctulata)--2 Chestnut-backed(Myrmeciza exsul)--20, 50, 51, 110, 111, 112, 116, 130, 133, 134, al Lunulated ( Gymnopithys lunulata)--28 Ocellated(Phaenostictus rncleannani)--18, 20, 27, 28, 42, 76, 78, 106, 110, 111, 112, 115, 116, 119, 120, 122 Rufous-throated ( Gy•nnopithys ru/igula)--33 Scale-backed( Hylophylax poecilonota)--97, 98 Spot-backed (Hylophylax naevia)--2, 93, 95 White-bellied ( Myrmeciza longipes)--116 Antbirds (Formicariidae)--74, 132 Antpitta, Streak-chested( Grailaria perspicillata)--116 Antshrike, Cinereous( Tharnnomanescaesius )--143 Dusky-throated ( Thamnomanes ardesiacus)--14 3 Slaw (Thamnophilus punctatusatrinuchus)--20, 48, 52, 54, 69, 76, 77, 78, 110, 112, 122, 127, 128, 129, 134, al Ant-Tanager, Red-crowned (Habia rubica)--76, 77, 78, 127, 128, 146 Red-throated (Habia/uscicauda)--64, 76, 77, 78, 100, 101, 127, 128, 143, 146 Antthrush, Black-faced (Formicarius analis)--116 Antvireo, Dot-crowned (Dysitharnnuspuncticeps) • 129, 132, a 1 Antwren, Brown-bellied ( Myrrnotherula gutturalis)--129 Checker-throated(Myrmotherula/ulviventris)--110, 113, 128, 129, 134, 142, 143, al Dot-winged (Microrhopias quixensis)--6, 7, 127, 130, 132, 134, 136, al Fulvous-bellied (see Checker-throated) Gray (Myrrnotherula rnenetriesii)--146 Long-winged (Myrmotherula longipennis)--146 Rufous-tailed (Myrrnotherula erythrura)--129 Stipple-throated (Myrrnotherula haernatonota)--129 White-eyed ( Myrmotherula leucophthalrna)--129 White-flanked (Myrmotherula axillaris)--78, 110, 126, 127, 128, 129, 133, 136, 142, 145, 146, al Apodidae--see Swifts Autornolus ochrolaernus--see Foliage-gleaner, Buff-throated Baryphthengusru/icapillt•ts--see Motmot, Great Rufous Becards (Pachyramphus spp., Platypsaris spp.)--132 Blackbirds (Icteridae)--135 Bluebird, Eastern ( Sialia sialis)•142 Bucconidae•see Puffbirds Bushtit (Psaltriparus minirnus)--138 Buteogallus anthracinus--see Hawk, Common Black Capella gallinago•see Snipe, Common Chickadees (Parus spp.)--135 Contopus spp.--see Pewees, Wood Cotingas (Cotingidae)--132, 140 Cuckoo, Squirrel (Piaya cayana)--110, 111, 113 Cyphorhinusphaeocephalus--see Wren, Song

• symbol: a • appendix number.

158 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 159

Dendrocincla /uliginosa--see Woodcreeper, Plain-brown Dendrocolaptes certhia--see Woodcreeper, Barred Dendrocolaptidae--see Woodcreepers Dendroica spp.--see Warblers Dysithamnus puncticeps--see Antvireo, Dot-crowned Empidonax virescens--see Flycatcher, Acadian Eucometis penicillata--see Tanager, Gray-headed Euphonia, Fulvous-vented(Euphonia /ulviventris)--130, al Flatbill, Olivaceous(Rhynchocyclus olivaceus)--132, al Flycatcher,Acadian (Empidonax virescens)--i 10, 133, al Bentbill (Oncostoma cinereigulare)--131, 135, 136, al Black-tailed (Myiobius atricaudus)--131, 138 Dusky-capped(Myiarchus tuberculi/er)--132, 133, 138, al Great Crested (Myiarchus crinitus)--133, al Ochre-bellied (Pipromorpha oleaginea)--131, al Olivaceous (see Dusky-capped) Ruddy-tailed (Terenotriccuserythrurus)--9, 110, 113, 131, al Sulphur-rumped( Myiobius sulphureipygeus)--131, 138 Yellow-margined ( Tolmomyias assimilis)--132, a 1 Flycatchers--see Tyrant Flycatchers Foliage-gleaner,Buff-throated ( Automolus ochrolaemus)--130 Forest-falcon, Barred ( M icrastur ru[icollis)--140 Forest-falcons(Micrastur spp.)--28, 140 Formicariidae--see Antbirds Formicarius analis see Antthrush, Black-faced Furnariidae--see Ovenbirds Glaucidium spp.--see Owls, Pygmy Glyphorynchusspirurus--see Woodcreeper,Wedge-billed Gnatcatcher, Tropical (Polioptila plumbea)--130, al Gnatwren, Long-billed (Ramphocaenus melanurus)--130, al Goldfinches (Spinus spp.)--79 Grailaria perspicillata--see Antpitta, Streak-chested Greenlet, Gray-headed (Hylophilus decurtatus)--126, 127, al Greenlets (Hylophilus spp.)--126, 127 Gymnopithys spp.--see Antbirds Habia spp.--see Ant-Tanagers Harpagus bidentatu•--see Kite, Double-toothed Hawk, Common Black (Buteogallusanthracinus)--20 Semiplumbeous( Leucopternissemiplumbea )--7 5, 84 Tiny (•lccipiter superciliosus)--140 White ( Leucopternis albicollis)--84 Hawks (Accipitridae, Falconidae)--140 Helmitheros vermivorus--see Warbler, Worm-eating Heterospingusrubri/rons--see Tanager, Sulphur-rumped Hummingbird, Hermit (Phaethornis spp.)-•28, 78 Hummingbirds(Trochilidae)--74 Hylocichla spp.--see Thrushes and Veery Hylophilus spp.--see Greenlets Hylophylax spp.--see Antbirds Icterids (Icteridae)--74, 135 160 ORNITHOLOGICAL MONOGRAPHS NO. 10

Kite, Double-toothed (Harpagus bidentatus)--20 Leafscraper,Scaly-throated (Sclerurus guatemalensis)--110, 113, 126, 138 Leafscrapers(Sclerurus spp.)--142 Leucopternis albicollis---see Hawk, White Malacoptila panamensis see Puffbird, White-whiskered Manacus vitellinus--see Manakin, Golden-collared Manakin, Blue-crowned ( Pipra coronata)--131 Golden-collared ( Manacus vitellinus)--131, al Red-capped (Pipra mentalis)--110, 131, al Manakins (Pipridae)--74, 131, 132, 139 Micrastur spp.--see Forest-falcons Microrhopias quixensis--see Antwren, Dot-winged Mniotilta varia-•see Warbler, Black-and-white Motmot, Great Rufous (Baryphthengusru/icapillus)--20, I I0, 113 Motmots (Momotidae)--118 Myiarchus spp.--see Flycatchers Myiobius spp.--see Flycatchers Myrmeciza spp.--see Antbirds Myrmotherula spp.--see Antwrens Oncostoma cinereigulare--see Flycatcher, Bentbill Oporornis formosus•see Warbler, Kentucky Ovenbirds (Furnariidae)--74, 130, 132 Owls, Pygmy (Glaucidium spp.)--140 Pachyramphus spp.--see Becards Paridac--see Titmice Parrots (Psittacidae)--74, 126, 132 Parulidae•see Warblers (Parulidae) Pewees, Wood (Contopus spp.)--133 Phaenostictus mcleannani--see Antbird, Ocellated Phaethornis spp.--see Hummingbirds, Hermit Piaya cayana--see Cuckoo, Squirrel Picidae--see Woodpeckers Pipra spp.--see Manakins Pipridae--see Manakins Pipromorpha oleaginea--see Flycatcher, Ochre-bellied Platypsaris spp.--see Becards Platyrinchus coronatus--see Spadebill, Golden-crowned Polioptila plumbea--see Gnatcatcher, Tropical Psaltriparus rainlinus--see Bushtit Psittacidae--see Parrots Puffbird, White-whiskered (Malacoptila panamensis)--20, 1 I0, 111, 113, al Puffbirds (Bucconidae)---74, 118 Ramphastidae--see Toucans Ramphastos sulfuratus--see Toucan, Keel-billed Ramphocaenusmelanurus--see Gnatwren, Long-billed Redstart, American (Setophaga ruticilla)--al Rhynchocyclus olivaceus--see Flatbill, Olivaceous $clerurus guatemalensis see Leafscraper, Scaly-throated Seiurus spp.--see Waterthrushes Setophagaruticilla--see Redstart, American 1972 WILLIS: BEHAVIOR OF SPOTTED ANTBIRDS 161

Sialia sialis see Bluebird, Eastern Snipe, Common ( Capella gallinago)---26 Spadebill, Golden-crowned (Platyrinchus coronatus)--132, al Spinus spp.--see Goldfinches Swifts (Apodidae)--126 Tachyphonusspp.--see Tanagers Tanager, Gray-headed (Eucornetispenicillata)---9, 20, 110, 111, 112 Plain-colored (Tangara inornata)--140, 141 Sulphur-rumped( Heterospingusrubri/rons )--128, al Tawny-crowned (Tachyphonusdelattrei)--128 White-shouldered (Tachyphonus luctuosus)--127, al Tanagers (Thraupidae)--74, 126 Tangara inornata--see Tanager, Plain-colored Terenotriccuserythrurus--see Flycatcher, Ruddy-tailed Thamnomanes spp.--see Antshrikes Thamnophiluspunctatus atrinuchus--see Antshrike, Slaty Thranpidaelsee Tanagers Thrush, Gray-cheeked (Hylocichla minima)--110, 111 Hermit ( Hy locichla guttata )--142 Swainson's( Hylocichla ustulata)--110, 111 Wood (Hylocichla mustelina)--133 Thrushes (Hylocichla spp.)--83, 110, 111, 113, 121, 122 Titmice (Paridae)--135 Tolmomyias assimilis see Flycatcher, Yellow-margined Toucan, Keel-billed ( Ramphastossul[uratus )--57 Toucans (Ramphastidae)--126, 132, 140, 145 Trochilidae•see Hummingbirds Troglodytidaelsee Wrens Trogon, Black-throated (Trogon ru/us)--20, 132, 145 Trogons (Trogonidae)--74, 132, 140 Tyrant Flycatcher (Tyrannidae)--74, 132, 133, 138 Veery (Hylocichla /uscescens)--110 Vireo, Red-eyed (Vireo olivaceus)--al Warbler, Bay-breasted(Dendroica castanea)--133, al Black-and-white (Mniotilta varia)--133, a 1 Canada (Wilsonia canadensis)--110, 111,113, 133, 134, 138, al Chestnut-sided(Dendroica pensylvanica)--133, al Kentucky (Oporornis/ormosus)--110, 133, 134, 135, al Worm-eating (Helmitheros vermivorus)--129 Warblers (Parulidae)--133, 135, 142 Waterthrushes (Seiurus spp.)--133, 142 Wilsonia canadensis--see Warbler, Canada Woodcreeper, Barred (Dendrocolaptes certhia)--102, 110, 112 Black-striped (Xiphorhynchus lachrymosus)--50, 130, al Buff-throated (Xiphorhynchusguttatus)--110, 122, 130, al Plain-brown (Dendrocincla /uliginosa)--18, 20, 27, 50, 110, 111, 112, 121, 122, 131, al Wedge-billed (Glyphorynchusspirurus)--128, 129, al Woodcreepers (Dendrocolaptidae)--I 33 Woodpeckers(Picidae)--74, 132, 145 162 ORNITHOLOGICAL MONOGRAPHS NO. 10

Wren, Song ( Cyphorhinusphaeocephalus )--128 Wrens (Troglodytidae)--147 Xenops, Plain (Xenops minutus)--129, al Xiphorhynchusspp.--see Woodcreepers ORNITHOLOGICAL MONOGRAPHS

No. 1. A Distributional Study of the Birds of British Honduras, by Stephen M. Russell. 195 pp., 2 color plates. 1964. Price $4.50 ($3.60 to AOU members). No. 2. A Comparative Study of Some Social Communication Patterns in the Pelecaniformes, by Gerard Frederick van Tets. 88 pp., 4 text figures. 1965. Price $2.00 ($1.60 to AOU members). No. 3. The Birds of Kentucky, by Robert M. Mengel. Cloth bound,xiv + 581 pp., 4 color plates plus text figures and vignettes. 1965. Price $10.00 ($8.00 to AOU members). No. 4. Evolution of Some Arctic Gulls (Larus): an Experimental Study of Isolating Mechanisms, by Neal Griffith Smith. 99 pp., 62 text figures. 1966. Price $2.50 ($2.00 to AOU members). No. 5. A ComparativeLife-history Study of Four Speciesof Woodpeckers, by Louise de Kiriline Lawrence. 156 pp., 33 text figures. 1967. Price $3.75 ($3.00 to AOU members). No. 6. Adaptations for Locomotionand Feeding in the Anhinga and the Double-crestedCormorant, by OscarT. Owre. 138 pp., 56 text figures. 1967. Price $3.50 ($2.80 to AOU members). No. 7. A Distributional Survey of the Birds of Honduras, by Burr L. Mortroe, Jr. 458 pp., 28 text figures, 2 color plates. 1968. Price $9.00 ($7.20 to AOU members). No. 8. An Approachto theStudy of EcologicalRelationships among Grdss- land Birds, by John A. Wiens. 93 pp., 30 text figures. 1969. Price $2.50 ($2.00 to AOU members). No. 9. Mating Systems,Sexual Dimorphism, and the Role of Male North AmericanPassefine Birds in the Nesting Cycle,by JaredVerner and Mary F. Willson. 76 pp. 1969. Price $2.50 ($2.00 to AOU members). No. I0. The Behaviorof SpottedAntbirds, by EdwinO. Willis,vi + 162pp., 3 colorplates, 27 text figures. 1972. Price $6.00 ($4.75 to AOU members).

Like all other AOU publications,Ornithological Monographs are shipped prepaid. Make checkspayable to "The AmericanOrnithologists' Union." For the convenienceof thosewho wish to maintain completesets of Ornithological Monographsand to receivenew numbersimmediately upon issue,standing orders will be accepted.

Orderfrom: Burt L. Monroe, Jr., Treasurer, American Ornithologists' Union, Box 23447,Anchorage, Kentucky 40223